)]}' { "log": [ { "commit": "c0ecd6388360d930440cc5554026818895199923", "tree": "1065fabd5fa7be633442ae09bc2a4d9779d55299", "parents": [ "183d46ff422ef9f3d755b6808ef3faa6d009ba3a", "5560a612c20d3daacbf5da7913deefa5c31742f4" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 01 11:30:15 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 01 11:30:15 2024 -0700" }, "message": "Merge tag \u0027pci-v6.11-fixes-1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/pci/pci\n\nPull PCI fixes from Bjorn Helgaas:\n\n - Fix a pci_intx() regression that caused driver reload to fail with\n \"Resources present before probing\" (Philipp Stanner)\n\n - Fix a pciehp regression that clobbered the upper bits of RAID status\n LEDs on NVMe devices behind an Intel VMD (Blazej Kucman)\n\n* tag \u0027pci-v6.11-fixes-1\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/pci/pci:\n PCI: pciehp: Retain Power Indicator bits for userspace indicators\n PCI: Fix devres regression in pci_intx()\n" }, { "commit": "5560a612c20d3daacbf5da7913deefa5c31742f4", "tree": "b0063ac82f5651ae1323782eed96f8752c4ca39e", "parents": [ "00f89ae4e759a7eef07e4188e1534af7dd2c7e9c" ], "author": { "name": "Blazej Kucman", "email": "blazej.kucman@intel.com", "time": "Mon Jul 22 16:14:40 2024 +0200" }, "committer": { "name": "Bjorn Helgaas", "email": "bhelgaas@google.com", "time": "Thu Aug 01 12:58:03 2024 -0500" }, "message": "PCI: pciehp: Retain Power Indicator bits for userspace indicators\n\nThe sysfs \"attention\" file normally controls the Slot Control Attention\nIndicator with 0 (off), 1 (on), 2 (blink) settings.\n\n576243b3f9ea (\"PCI: pciehp: Allow exclusive userspace control of\nindicators\") added pciehp_set_raw_indicator_status() to allow userspace to\ndirectly control all four bits in both the Attention Indicator and the\nPower Indicator fields via the \"attention\" file.\n\nThis is used on Intel VMD bridges so utilities like \"ledmon\" can use sysfs\n\"attention\" to control up to 16 indicators for NVMe device RAID status.\n\nabaaac4845a0 (\"PCI: hotplug: Use FIELD_GET/PREP()\") broke this by masking\nthe sysfs data with PCI_EXP_SLTCTL_AIC, which discards the upper two bits\nintended for the Power Indicator Control field (PCI_EXP_SLTCTL_PIC).\n\nFor NVMe devices behind an Intel VMD, ledmon settings that use the\nPCI_EXP_SLTCTL_PIC bits, i.e., ATTENTION_REBUILD (0x5), ATTENTION_LOCATE\n(0x7), ATTENTION_FAILURE (0xD), ATTENTION_OFF (0xF), no longer worked\ncorrectly.\n\nMask with PCI_EXP_SLTCTL_AIC | PCI_EXP_SLTCTL_PIC to retain both the\nAttention Indicator and the Power Indicator bits.\n\nFixes: abaaac4845a0 (\"PCI: hotplug: Use FIELD_GET/PREP()\")\nLink: https://lore.kernel.org/r/20240722141440.7210-1-blazej.kucman@intel.com\nSigned-off-by: Blazej Kucman \u003cblazej.kucman@intel.com\u003e\n[bhelgaas: commit log]\nSigned-off-by: Bjorn Helgaas \u003cbhelgaas@google.com\u003e\nCc: stable@vger.kernel.org\t# v6.7+\n" }, { "commit": "00f89ae4e759a7eef07e4188e1534af7dd2c7e9c", "tree": "cf9339bed0f280f7fb2e4843ded6137b36c39c2a", "parents": [ "8400291e289ee6b2bf9779ff1c83a291501f017b" ], "author": { "name": "Philipp Stanner", "email": "pstanner@redhat.com", "time": "Thu Jul 25 14:07:30 2024 +0200" }, "committer": { "name": "Bjorn Helgaas", "email": "bhelgaas@google.com", "time": "Thu Aug 01 12:56:06 2024 -0500" }, "message": "PCI: Fix devres regression in pci_intx()\n\npci_intx() becomes managed if pcim_enable_device() has been called in\nadvance. Commit 25216afc9db5 (\"PCI: Add managed pcim_intx()\") changed this\nbehavior so that pci_intx() always leads to creation of a separate device\nresource for itself, whereas earlier, a shared resource was used for all\nPCI devres operations.\n\nUnfortunately, pci_intx() seems to be used in some drivers\u0027 remove() paths;\nin the managed case this causes a device resource to be created on driver\ndetach, which causes .probe() to fail if the driver is reloaded:\n\n pci 0000:00:1f.2: Resources present before probing\n\nFix the regression by only redirecting pci_intx() to its managed twin\npcim_intx() if the pci_command changes.\n\nLink: https://lore.kernel.org/r/20240725120729.59788-2-pstanner@redhat.com\nFixes: 25216afc9db5 (\"PCI: Add managed pcim_intx()\")\nReported-by: Damien Le Moal \u003cdlemoal@kernel.org\u003e\nCloses: https://lore.kernel.org/all/b8f4ba97-84fc-4b7e-ba1a-99de2d9f0118@kernel.org/\nSigned-off-by: Philipp Stanner \u003cpstanner@redhat.com\u003e\n[bhelgaas: add error message to commit log]\nSigned-off-by: Bjorn Helgaas \u003cbhelgaas@google.com\u003e\nTested-by: Damien Le Moal \u003cdlemoal@kernel.org\u003e\n" }, { "commit": "183d46ff422ef9f3d755b6808ef3faa6d009ba3a", "tree": "8405b88ef447994f3eafd40e8e3aa5f47d3fa33f", "parents": [ "21b136cc63d2a9ddd60d4699552b69c214b32964", "25010bfdf8bbedc64c5c04d18f846412f5367d26" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 01 09:42:09 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Aug 01 09:42:09 2024 -0700" }, "message": "Merge tag \u0027net-6.11-rc2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net\n\nPull networking fixes from Paolo Abeni:\n \"Including fixes from wireless, bleutooth, BPF and netfilter.\n\n Current release - regressions:\n\n - core: drop bad gso csum_start and offset in virtio_net_hdr\n\n - wifi: mt76: fix null pointer access in mt792x_mac_link_bss_remove\n\n - eth: tun: add missing bpf_net_ctx_clear() in do_xdp_generic()\n\n - phy: aquantia: only poll GLOBAL_CFG regs on aqr113, aqr113c and\n aqr115c\n\n Current release - new code bugs:\n\n - smc: prevent UAF in inet_create()\n\n - bluetooth: btmtk: fix kernel crash when entering btmtk_usb_suspend\n\n - eth: bnxt: reject unsupported hash functions\n\n Previous releases - regressions:\n\n - sched: act_ct: take care of padding in struct zones_ht_key\n\n - netfilter: fix null-ptr-deref in iptable_nat_table_init().\n\n - tcp: adjust clamping window for applications specifying SO_RCVBUF\n\n Previous releases - always broken:\n\n - ethtool: rss: small fixes to spec and GET\n\n - mptcp:\n - fix signal endpoint re-add\n - pm: fix backup support in signal endpoints\n\n - wifi: ath12k: fix soft lockup on suspend\n\n - eth: bnxt_en: fix RSS logic in __bnxt_reserve_rings()\n\n - eth: ice: fix AF_XDP ZC timeout and concurrency issues\n\n - eth: mlx5:\n - fix missing lock on sync reset reload\n - fix error handling in irq_pool_request_irq\"\n\n* tag \u0027net-6.11-rc2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (76 commits)\n mptcp: fix duplicate data handling\n mptcp: fix bad RCVPRUNED mib accounting\n ipv6: fix ndisc_is_useropt() handling for PIO\n igc: Fix double reset adapter triggered from a single taprio cmd\n net: MAINTAINERS: Demote Qualcomm IPA to \"maintained\"\n net: wan: fsl_qmc_hdlc: Discard received CRC\n net: wan: fsl_qmc_hdlc: Convert carrier_lock spinlock to a mutex\n net/mlx5e: Add a check for the return value from mlx5_port_set_eth_ptys\n net/mlx5e: Fix CT entry update leaks of modify header context\n net/mlx5e: Require mlx5 tc classifier action support for IPsec prio capability\n net/mlx5: Fix missing lock on sync reset reload\n net/mlx5: Lag, don\u0027t use the hardcoded value of the first port\n net/mlx5: DR, Fix \u0027stack guard page was hit\u0027 error in dr_rule\n net/mlx5: Fix error handling in irq_pool_request_irq\n net/mlx5: Always drain health in shutdown callback\n net: Add skbuff.h to MAINTAINERS\n r8169: don\u0027t increment tx_dropped in case of NETDEV_TX_BUSY\n netfilter: iptables: Fix potential null-ptr-deref in ip6table_nat_table_init().\n netfilter: iptables: Fix null-ptr-deref in iptable_nat_table_init().\n net: drop bad gso csum_start and offset in virtio_net_hdr\n ...\n" }, { "commit": "25010bfdf8bbedc64c5c04d18f846412f5367d26", "tree": "ce89414485cd2c82a99bc5c5e65d711bea4889b9", "parents": [ "2b4a32daa6de7ff4c978c07878880ab2bb475932", "68cc924729ffcfe90d0383177192030a9aeb2ee4" ], "author": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Thu Aug 01 12:30:15 2024 +0200" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Thu Aug 01 12:30:16 2024 +0200" }, "message": "Merge branch \u0027mptcp-fix-duplicate-data-handling\u0027\n\nMatthieu Baerts says:\n\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\nmptcp: fix duplicate data handling\n\nIn some cases, the subflow-level\u0027s copied_seq counter was incorrectly\nincreased, leading to an unexpected subflow reset.\n\nPatch 1/2 fixes the RCVPRUNED MIB counter that was attached to the wrong\nevent since its introduction in v5.14, backported to v5.11.\n\nPatch 2/2 fixes the copied_seq counter issues, is present since v5.10.\n\nSigned-off-by: Matthieu Baerts (NGI0) \u003cmatttbe@kernel.org\u003e\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\nLink: https://patch.msgid.link/20240731-upstream-net-20240731-mptcp-dup-data-v1-0-bde833fa628a@kernel.org\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n" }, { "commit": "68cc924729ffcfe90d0383177192030a9aeb2ee4", "tree": "ce89414485cd2c82a99bc5c5e65d711bea4889b9", "parents": [ "0a567c2a10033bf04ed618368d179bce6977984b" ], "author": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Wed Jul 31 12:10:15 2024 +0200" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Thu Aug 01 12:30:13 2024 +0200" }, "message": "mptcp: fix duplicate data handling\n\nWhen a subflow receives and discards duplicate data, the mptcp\nstack assumes that the consumed offset inside the current skb is\nzero.\n\nWith multiple subflows receiving data simultaneously such assertion\ndoes not held true. As a result the subflow-level copied_seq will\nbe incorrectly increased and later on the same subflow will observe\na bad mapping, leading to subflow reset.\n\nAddress the issue taking into account the skb consumed offset in\nmptcp_subflow_discard_data().\n\nFixes: 04e4cd4f7ca4 (\"mptcp: cleanup mptcp_subflow_discard_data()\")\nCc: stable@vger.kernel.org\nLink: https://github.com/multipath-tcp/mptcp_net-next/issues/501\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\nReviewed-by: Mat Martineau \u003cmartineau@kernel.org\u003e\nSigned-off-by: Matthieu Baerts (NGI0) \u003cmatttbe@kernel.org\u003e\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n\n" }, { "commit": "0a567c2a10033bf04ed618368d179bce6977984b", "tree": "02dac07cb509fb83eee5b156023b27fac3bc8cac", "parents": [ "2b4a32daa6de7ff4c978c07878880ab2bb475932" ], "author": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Wed Jul 31 12:10:14 2024 +0200" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Thu Aug 01 12:30:13 2024 +0200" }, "message": "mptcp: fix bad RCVPRUNED mib accounting\n\nSince its introduction, the mentioned MIB accounted for the wrong\nevent: wake-up being skipped as not-needed on some edge condition\ninstead of incoming skb being dropped after landing in the (subflow)\nreceive queue.\n\nMove the increment in the correct location.\n\nFixes: ce599c516386 (\"mptcp: properly account bulk freed memory\")\nCc: stable@vger.kernel.org\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\nReviewed-by: Mat Martineau \u003cmartineau@kernel.org\u003e\nSigned-off-by: Matthieu Baerts (NGI0) \u003cmatttbe@kernel.org\u003e\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n\n" }, { "commit": "2b4a32daa6de7ff4c978c07878880ab2bb475932", "tree": "8dbfca8ab22ae7b25f3577acfa48201bce70cbc2", "parents": [ "a46c68debf3be3a477a69ccbf0a1d050df841676", "c22921df777de5606f1047b1345b8d22ef1c0b34" ], "author": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Thu Aug 01 12:08:28 2024 +0200" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Thu Aug 01 12:08:28 2024 +0200" }, "message": "Merge tag \u0027nf-24-07-31\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf\n\nPablo Neira Ayuso says:\n\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\nNetfilter fixes for net\n\nThe following patchset contains Netfilter fixes for net:\n\nFix a possible null-ptr-deref sometimes triggered by iptables-restore at\nboot time. Register iptables {ipv4,ipv6} nat table pernet in first place\nto fix this issue. Patch #1 and #2 from Kuniyuki Iwashima.\n\nnetfilter pull request 24-07-31\n\n* tag \u0027nf-24-07-31\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf:\n netfilter: iptables: Fix potential null-ptr-deref in ip6table_nat_table_init().\n netfilter: iptables: Fix null-ptr-deref in iptable_nat_table_init().\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\nLink: https://patch.msgid.link/20240731213046.6194-1-pablo@netfilter.org\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n" }, { "commit": "a46c68debf3be3a477a69ccbf0a1d050df841676", "tree": "fed89c57135e389343d10e0c6e28c62fcb944b3c", "parents": [ "b9e7fc0aeda79031a101610b2fcb12bf031056e9" ], "author": { "name": "Maciej Żenczykowski", "email": "maze@google.com", "time": "Mon Jul 29 17:17:48 2024 -0700" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Thu Aug 01 11:40:29 2024 +0200" }, "message": "ipv6: fix ndisc_is_useropt() handling for PIO\n\nThe current logic only works if the PIO is between two\nother ND user options. This fixes it so that the PIO\ncan also be either before or after other ND user options\n(for example the first or last option in the RA).\n\nside note: there\u0027s actually Android tests verifying\na portion of the old broken behaviour, so:\n https://android-review.googlesource.com/c/kernel/tests/+/3196704\nfixes those up.\n\nCc: Jen Linkova \u003cfurry@google.com\u003e\nCc: Lorenzo Colitti \u003clorenzo@google.com\u003e\nCc: Patrick Rohr \u003cprohr@google.com\u003e\nCc: David Ahern \u003cdsahern@kernel.org\u003e\nCc: YOSHIFUJI Hideaki / 吉藤英明 \u003cyoshfuji@linux-ipv6.org\u003e\nCc: Jakub Kicinski \u003ckuba@kernel.org\u003e\nSigned-off-by: Maciej Żenczykowski \u003cmaze@google.com\u003e\nFixes: 048c796beb6e (\"ipv6: adjust ndisc_is_useropt() to also return true for PIO\")\nLink: https://patch.msgid.link/20240730001748.147636-1-maze@google.com\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n\n" }, { "commit": "b9e7fc0aeda79031a101610b2fcb12bf031056e9", "tree": "1aeb60838ffb9982261275de639267713d40586a", "parents": [ "4efce726e0cbc723178eea5b944e13775f628ecc" ], "author": { "name": "Faizal Rahim", "email": "faizal.abdul.rahim@linux.intel.com", "time": "Tue Jul 30 10:33:02 2024 -0700" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Wed Jul 31 18:56:23 2024 -0700" }, "message": "igc: Fix double reset adapter triggered from a single taprio cmd\n\nFollowing the implementation of \"igc: Add TransmissionOverrun counter\"\npatch, when a taprio command is triggered by user, igc processes two\ncommands: TAPRIO_CMD_REPLACE followed by TAPRIO_CMD_STATS. However, both\ncommands unconditionally pass through igc_tsn_offload_apply() which\nevaluates and triggers reset adapter. The double reset causes issues in\nthe calculation of adapter-\u003eqbv_count in igc.\n\nTAPRIO_CMD_REPLACE command is expected to reset the adapter since it\nactivates qbv. It\u0027s unexpected for TAPRIO_CMD_STATS to do the same\nbecause it doesn\u0027t configure any driver-specific TSN settings. So, the\nevaluation in igc_tsn_offload_apply() isn\u0027t needed for TAPRIO_CMD_STATS.\n\nTo address this, commands parsing are relocated to\nigc_tsn_enable_qbv_scheduling(). Commands that don\u0027t require an adapter\nreset will exit after processing, thus avoiding igc_tsn_offload_apply().\n\nFixes: d3750076d464 (\"igc: Add TransmissionOverrun counter\")\nSigned-off-by: Faizal Rahim \u003cfaizal.abdul.rahim@linux.intel.com\u003e\nAcked-by: Vinicius Costa Gomes \u003cvinicius.gomes@intel.com\u003e\nReviewed-by: Vladimir Oltean \u003cvladimir.oltean@nxp.com\u003e\nTested-by: Mor Bar-Gabay \u003cmorx.bar.gabay@intel.com\u003e\nSigned-off-by: Tony Nguyen \u003canthony.l.nguyen@intel.com\u003e\nLink: https://patch.msgid.link/20240730173304.865479-1-anthony.l.nguyen@intel.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "4efce726e0cbc723178eea5b944e13775f628ecc", "tree": "4cb885b44ee00959b5bbefcc2713fc39f25868a4", "parents": [ "e549360069b4a57e111b8222fc072f3c7c1688ab" ], "author": { "name": "Krzysztof Kozlowski", "email": "krzysztof.kozlowski@linaro.org", "time": "Tue Jul 30 12:40:16 2024 +0200" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Wed Jul 31 18:10:28 2024 -0700" }, "message": "net: MAINTAINERS: Demote Qualcomm IPA to \"maintained\"\n\nTo the best of my knowledge, Alex Elder is not being paid to support\nQualcomm IPA networking drivers, so drop the status from \"supported\" to\n\"maintained\".\n\nSigned-off-by: Krzysztof Kozlowski \u003ckrzysztof.kozlowski@linaro.org\u003e\nAcked-by: Alex Elder \u003celder@kernel.org\u003e\nLink: https://patch.msgid.link/20240730104016.22103-1-krzysztof.kozlowski@linaro.org\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "e549360069b4a57e111b8222fc072f3c7c1688ab", "tree": "d010976d949181d16547e0d330329dfc2b3c8649", "parents": [ "c4d6a347ba7babdf9d90a0eb24048c266cae0532" ], "author": { "name": "Herve Codina", "email": "herve.codina@bootlin.com", "time": "Tue Jul 30 08:31:33 2024 +0200" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Wed Jul 31 18:09:08 2024 -0700" }, "message": "net: wan: fsl_qmc_hdlc: Discard received CRC\n\nReceived frame from QMC contains the CRC.\nUpper layers don\u0027t need this CRC and tcpdump mentioned trailing junk\ndata due to this CRC presence.\n\nAs some other HDLC driver, simply discard this CRC.\n\nFixes: d0f2258e79fd (\"net: wan: Add support for QMC HDLC\")\nCc: stable@vger.kernel.org\nSigned-off-by: Herve Codina \u003cherve.codina@bootlin.com\u003e\nReviewed-by: Simon Horman \u003chorms@kernel.org\u003e\nLink: https://patch.msgid.link/20240730063133.179598-1-herve.codina@bootlin.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "c4d6a347ba7babdf9d90a0eb24048c266cae0532", "tree": "4882ab97fb7962266e18a2ed0f1a05f44fc53e35", "parents": [ "fe16667f2959ec89488a872efbd9782bed250d10" ], "author": { "name": "Herve Codina", "email": "herve.codina@bootlin.com", "time": "Tue Jul 30 08:31:04 2024 +0200" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Wed Jul 31 18:06:24 2024 -0700" }, "message": "net: wan: fsl_qmc_hdlc: Convert carrier_lock spinlock to a mutex\n\nThe carrier_lock spinlock protects the carrier detection. While it is\nheld, framer_get_status() is called which in turn takes a mutex.\nThis is not correct and can lead to a deadlock.\n\nA run with PROVE_LOCKING enabled detected the issue:\n [ BUG: Invalid wait context ]\n ...\n c204ddbc (\u0026framer-\u003emutex){+.+.}-{3:3}, at: framer_get_status+0x40/0x78\n other info that might help us debug this:\n context-{4:4}\n 2 locks held by ifconfig/146:\n #0: c0926a38 (rtnl_mutex){+.+.}-{3:3}, at: devinet_ioctl+0x12c/0x664\n #1: c2006a40 (\u0026qmc_hdlc-\u003ecarrier_lock){....}-{2:2}, at: qmc_hdlc_framer_set_carrier+0x30/0x98\n\nAvoid the spinlock usage and convert carrier_lock to a mutex.\n\nFixes: 54762918ca85 (\"net: wan: fsl_qmc_hdlc: Add framer support\")\nCc: stable@vger.kernel.org\nSigned-off-by: Herve Codina \u003cherve.codina@bootlin.com\u003e\nReviewed-by: Simon Horman \u003chorms@kernel.org\u003e\nLink: https://patch.msgid.link/20240730063104.179553-1-herve.codina@bootlin.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "fe16667f2959ec89488a872efbd9782bed250d10", "tree": "e55f67ba4f7698e9ca4eeecb36d7a4084d54a32e", "parents": [ "8f73ef82985890e484efaed816b172fdf35c87aa", "3f8e82a020a5c22f9b791f4ac499b8e18007fbda" ], "author": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Wed Jul 31 18:04:53 2024 -0700" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Wed Jul 31 18:04:53 2024 -0700" }, "message": "Merge branch \u0027mlx5-misc-fixes-2024-07-30\u0027\n\nTariq Toukan says:\n\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\nmlx5 misc fixes 2024-07-30\n\nThis patchset provides misc bug fixes from the team to the mlx5 core and\nEth drivers.\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\nLink: https://patch.msgid.link/20240730061638.1831002-1-tariqt@nvidia.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "3f8e82a020a5c22f9b791f4ac499b8e18007fbda", "tree": "e55f67ba4f7698e9ca4eeecb36d7a4084d54a32e", "parents": [ "025f2b85a5e5a46df14ecf162c3c80a957a36d0b" ], "author": { "name": "Shahar Shitrit", "email": "shshitrit@nvidia.com", "time": "Tue Jul 30 09:16:37 2024 +0300" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Wed Jul 31 18:04:51 2024 -0700" }, "message": "net/mlx5e: Add a check for the return value from mlx5_port_set_eth_ptys\n\nSince the documentation for mlx5_toggle_port_link states that it should\nonly be used after setting the port register, we add a check for the\nreturn value from mlx5_port_set_eth_ptys to ensure the register was\nsuccessfully set before calling it.\n\nFixes: 667daedaecd1 (\"net/mlx5e: Toggle link only after modifying port parameters\")\nSigned-off-by: Shahar Shitrit \u003cshshitrit@nvidia.com\u003e\nReviewed-by: Carolina Jubran \u003ccjubran@nvidia.com\u003e\nSigned-off-by: Tariq Toukan \u003ctariqt@nvidia.com\u003e\nReviewed-by: Wojciech Drewek \u003cwojciech.drewek@intel.com\u003e\nLink: https://patch.msgid.link/20240730061638.1831002-9-tariqt@nvidia.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "025f2b85a5e5a46df14ecf162c3c80a957a36d0b", "tree": "8d3a86c18fe88ff0f813e23eca54dcf687181ec8", "parents": [ "06827e27fdcd197557be72b2229dbd362303794f" ], "author": { "name": "Chris Mi", "email": "cmi@nvidia.com", "time": "Tue Jul 30 09:16:36 2024 +0300" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Wed Jul 31 18:04:51 2024 -0700" }, "message": "net/mlx5e: Fix CT entry update leaks of modify header context\n\nThe cited commit allocates a new modify header to replace the old\none when updating CT entry. But if failed to allocate a new one, eg.\nexceed the max number firmware can support, modify header will be\nan error pointer that will trigger a panic when deallocating it. And\nthe old modify header point is copied to old attr. When the old\nattr is freed, the old modify header is lost.\n\nFix it by restoring the old attr to attr when failed to allocate a\nnew modify header context. So when the CT entry is freed, the right\nmodify header context will be freed. And the panic of accessing\nerror pointer is also fixed.\n\nFixes: 94ceffb48eac (\"net/mlx5e: Implement CT entry update\")\nSigned-off-by: Chris Mi \u003ccmi@nvidia.com\u003e\nSigned-off-by: Tariq Toukan \u003ctariqt@nvidia.com\u003e\nReviewed-by: Wojciech Drewek \u003cwojciech.drewek@intel.com\u003e\nLink: https://patch.msgid.link/20240730061638.1831002-8-tariqt@nvidia.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "06827e27fdcd197557be72b2229dbd362303794f", "tree": "f16da300c3ee37803c8dc91c2f1f3b34c7453f82", "parents": [ "572f9caa9e7295f8c8822e4122c7ae8f1c412ff9" ], "author": { "name": "Rahul Rameshbabu", "email": "rrameshbabu@nvidia.com", "time": "Tue Jul 30 09:16:35 2024 +0300" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Wed Jul 31 18:04:51 2024 -0700" }, "message": "net/mlx5e: Require mlx5 tc classifier action support for IPsec prio capability\n\nRequire mlx5 classifier action support when creating IPSec chains in\noffload path. MLX5_IPSEC_CAP_PRIO should only be set if CONFIG_MLX5_CLS_ACT\nis enabled. If CONFIG_MLX5_CLS_ACT\u003dn and MLX5_IPSEC_CAP_PRIO is set,\nconfiguring IPsec offload will fail due to the mlxx5 ipsec chain rules\nfailing to be created due to lack of classifier action support.\n\nFixes: fa5aa2f89073 (\"net/mlx5e: Use chains for IPsec policy priority offload\")\nSigned-off-by: Rahul Rameshbabu \u003crrameshbabu@nvidia.com\u003e\nReviewed-by: Leon Romanovsky \u003cleonro@nvidia.com\u003e\nSigned-off-by: Tariq Toukan \u003ctariqt@nvidia.com\u003e\nReviewed-by: Wojciech Drewek \u003cwojciech.drewek@intel.com\u003e\nLink: https://patch.msgid.link/20240730061638.1831002-7-tariqt@nvidia.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "572f9caa9e7295f8c8822e4122c7ae8f1c412ff9", "tree": "f087e75613b8d0cee85a8419333dd948bf83c8ce", "parents": [ "3fda84dc090390573cfbd0b1d70372663315de21" ], "author": { "name": "Moshe Shemesh", "email": "moshe@nvidia.com", "time": "Tue Jul 30 09:16:34 2024 +0300" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Wed Jul 31 18:04:51 2024 -0700" }, "message": "net/mlx5: Fix missing lock on sync reset reload\n\nOn sync reset reload work, when remote host updates devlink on reload\nactions performed on that host, it misses taking devlink lock before\ncalling devlink_remote_reload_actions_performed() which results in\ntriggering lock assert like the following:\n\nWARNING: CPU: 4 PID: 1164 at net/devlink/core.c:261 devl_assert_locked+0x3e/0x50\n…\n CPU: 4 PID: 1164 Comm: kworker/u96:6 Tainted: G S W 6.10.0-rc2+ #116\n Hardware name: Supermicro SYS-2028TP-DECTR/X10DRT-PT, BIOS 2.0 12/18/2015\n Workqueue: mlx5_fw_reset_events mlx5_sync_reset_reload_work [mlx5_core]\n RIP: 0010:devl_assert_locked+0x3e/0x50\n…\n Call Trace:\n \u003cTASK\u003e\n ? __warn+0xa4/0x210\n ? devl_assert_locked+0x3e/0x50\n ? report_bug+0x160/0x280\n ? handle_bug+0x3f/0x80\n ? exc_invalid_op+0x17/0x40\n ? asm_exc_invalid_op+0x1a/0x20\n ? devl_assert_locked+0x3e/0x50\n devlink_notify+0x88/0x2b0\n ? mlx5_attach_device+0x20c/0x230 [mlx5_core]\n ? __pfx_devlink_notify+0x10/0x10\n ? process_one_work+0x4b6/0xbb0\n process_one_work+0x4b6/0xbb0\n[…]\n\nFixes: 84a433a40d0e (\"net/mlx5: Lock mlx5 devlink reload callbacks\")\nSigned-off-by: Moshe Shemesh \u003cmoshe@nvidia.com\u003e\nReviewed-by: Maor Gottlieb \u003cmaorg@nvidia.com\u003e\nSigned-off-by: Tariq Toukan \u003ctariqt@nvidia.com\u003e\nReviewed-by: Wojciech Drewek \u003cwojciech.drewek@intel.com\u003e\nLink: https://patch.msgid.link/20240730061638.1831002-6-tariqt@nvidia.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "3fda84dc090390573cfbd0b1d70372663315de21", "tree": "6d3cdd45abb91ba6c1bb90e7f125ed0d3dcdf1fc", "parents": [ "94a3ad6c081381fa9ee523781789802b4ed00faf" ], "author": { "name": "Mark Bloch", "email": "mbloch@nvidia.com", "time": "Tue Jul 30 09:16:33 2024 +0300" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Wed Jul 31 18:04:51 2024 -0700" }, "message": "net/mlx5: Lag, don\u0027t use the hardcoded value of the first port\n\nThe cited commit didn\u0027t change the body of the loop as it should.\nIt shouldn\u0027t be using MLX5_LAG_P1.\n\nFixes: 7e978e7714d6 (\"net/mlx5: Lag, use actual number of lag ports\")\nSigned-off-by: Mark Bloch \u003cmbloch@nvidia.com\u003e\nSigned-off-by: Tariq Toukan \u003ctariqt@nvidia.com\u003e\nReviewed-by: Wojciech Drewek \u003cwojciech.drewek@intel.com\u003e\nLink: https://patch.msgid.link/20240730061638.1831002-5-tariqt@nvidia.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "94a3ad6c081381fa9ee523781789802b4ed00faf", "tree": "141fb0b767ef9fe01d3db8e12425ca724ed573d1", "parents": [ "a4557b0b57c40871ff00da4f623cf79211e052f3" ], "author": { "name": "Yevgeny Kliteynik", "email": "kliteyn@nvidia.com", "time": "Tue Jul 30 09:16:32 2024 +0300" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Wed Jul 31 18:04:51 2024 -0700" }, "message": "net/mlx5: DR, Fix \u0027stack guard page was hit\u0027 error in dr_rule\n\nThis patch reduces the size of hw_ste_arr_optimized array that is\nallocated on stack from 640 bytes (5 match STEs + 5 action STES)\nto 448 bytes (2 match STEs + 5 action STES).\nThis fixes the \u0027stack guard page was hit\u0027 issue, while still fitting\nmajority of the usecases (up to 2 match STEs).\n\nSigned-off-by: Yevgeny Kliteynik \u003ckliteyn@nvidia.com\u003e\nReviewed-by: Alex Vesker \u003cvalex@nvidia.com\u003e\nSigned-off-by: Tariq Toukan \u003ctariqt@nvidia.com\u003e\nReviewed-by: Wojciech Drewek \u003cwojciech.drewek@intel.com\u003e\nLink: https://patch.msgid.link/20240730061638.1831002-4-tariqt@nvidia.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "a4557b0b57c40871ff00da4f623cf79211e052f3", "tree": "958fef7e9939899b76e9b11acbdf019d4a608c83", "parents": [ "1b75da22ed1e6171e261bc9265370162553d5393" ], "author": { "name": "Shay Drory", "email": "shayd@nvidia.com", "time": "Tue Jul 30 09:16:31 2024 +0300" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Wed Jul 31 18:04:51 2024 -0700" }, "message": "net/mlx5: Fix error handling in irq_pool_request_irq\n\nIn case mlx5_irq_alloc fails, the previously allocated index remains\nin the XArray, which could lead to inconsistencies.\n\nFix it by adding error handling that erases the allocated index\nfrom the XArray if mlx5_irq_alloc returns an error.\n\nFixes: c36326d38d93 (\"net/mlx5: Round-Robin EQs over IRQs\")\nSigned-off-by: Shay Drory \u003cshayd@nvidia.com\u003e\nReviewed-by: Maher Sanalla \u003cmsanalla@nvidia.com\u003e\nSigned-off-by: Tariq Toukan \u003ctariqt@nvidia.com\u003e\nReviewed-by: Wojciech Drewek \u003cwojciech.drewek@intel.com\u003e\nLink: https://patch.msgid.link/20240730061638.1831002-3-tariqt@nvidia.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "1b75da22ed1e6171e261bc9265370162553d5393", "tree": "817ab84112580e09051bd543b499864e462515f1", "parents": [ "8f73ef82985890e484efaed816b172fdf35c87aa" ], "author": { "name": "Shay Drory", "email": "shayd@nvidia.com", "time": "Tue Jul 30 09:16:30 2024 +0300" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Wed Jul 31 18:04:50 2024 -0700" }, "message": "net/mlx5: Always drain health in shutdown callback\n\nThere is no point in recovery during device shutdown. if health\nwork started need to wait for it to avoid races and NULL pointer\naccess.\n\nHence, drain health WQ on shutdown callback.\n\nFixes: 1958fc2f0712 (\"net/mlx5: SF, Add auxiliary device driver\")\nFixes: d2aa060d40fa (\"net/mlx5: Cancel health poll before sending panic teardown command\")\nSigned-off-by: Shay Drory \u003cshayd@nvidia.com\u003e\nReviewed-by: Moshe Shemesh \u003cmoshe@nvidia.com\u003e\nSigned-off-by: Tariq Toukan \u003ctariqt@nvidia.com\u003e\nReviewed-by: Wojciech Drewek \u003cwojciech.drewek@intel.com\u003e\nLink: https://patch.msgid.link/20240730061638.1831002-2-tariqt@nvidia.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "8f73ef82985890e484efaed816b172fdf35c87aa", "tree": "9aba225e14c97f7022487beebd3161b15a4731e0", "parents": [ "d516b187a9cc2e842030dd005be2735db3e8f395" ], "author": { "name": "Breno Leitao", "email": "leitao@debian.org", "time": "Tue Jul 30 09:14:03 2024 -0700" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Wed Jul 31 17:53:37 2024 -0700" }, "message": "net: Add skbuff.h to MAINTAINERS\n\nThe network maintainers need to be copied if the skbuff.h is touched.\n\nThis also helps git-send-email to figure out the proper maintainers when\ntouching the file.\n\nSigned-off-by: Breno Leitao \u003cleitao@debian.org\u003e\nReviewed-by: Simon Horman \u003chorms@kernel.org\u003e\nLink: https://patch.msgid.link/20240730161404.2028175-1-leitao@debian.org\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "d516b187a9cc2e842030dd005be2735db3e8f395", "tree": "c11c1d62f9f2cab399f7e0c52f8a2ab67b8fb806", "parents": [ "601df205896da88e654d7c97d4579fa4af0b4adf" ], "author": { "name": "Heiner Kallweit", "email": "hkallweit1@gmail.com", "time": "Tue Jul 30 21:51:52 2024 +0200" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Wed Jul 31 17:52:46 2024 -0700" }, "message": "r8169: don\u0027t increment tx_dropped in case of NETDEV_TX_BUSY\n\nThe skb isn\u0027t consumed in case of NETDEV_TX_BUSY, therefore don\u0027t\nincrement the tx_dropped counter.\n\nFixes: 188f4af04618 (\"r8169: use NETDEV_TX_{BUSY/OK}\")\nCc: stable@vger.kernel.org\nSuggested-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\nSigned-off-by: Heiner Kallweit \u003chkallweit1@gmail.com\u003e\nReviewed-by: Wojciech Drewek \u003cwojciech.drewek@intel.com\u003e\nLink: https://patch.msgid.link/bbba9c48-8bac-4932-9aa1-d2ed63bc9433@gmail.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "601df205896da88e654d7c97d4579fa4af0b4adf", "tree": "33ff821d073b60a589ff9dbf7289291845d328c1", "parents": [ "0bf50cead4c4710d9f704778c32ab8af47ddf070", "7764b9622db4382b2797b54a70f292c8da6ef417" ], "author": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Wed Jul 31 17:49:00 2024 -0700" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Wed Jul 31 17:49:00 2024 -0700" }, "message": "Merge tag \u0027for-netdev\u0027 of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf\n\nDaniel Borkmann says:\n\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\npull-request: bpf 2024-07-31\n\nWe\u0027ve added 2 non-merge commits during the last 2 day(s) which contain\na total of 2 files changed, 2 insertions(+), 2 deletions(-).\n\nThe main changes are:\n\n1) Fix BPF selftest build after tree sync with regards to a _GNU_SOURCE\n macro redefined compilation error, from Stanislav Fomichev.\n\n2) Fix a wrong test in the ASSERT_OK() check in uprobe_syscall BPF selftest,\n from Jiri Olsa.\n\n* tag \u0027for-netdev\u0027 of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf:\n bpf/selftests: Fix ASSERT_OK condition check in uprobe_syscall test\n selftests/bpf: Filter out _GNU_SOURCE when compiling test_cpp\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\nLink: https://patch.msgid.link/20240731115706.19677-1-daniel@iogearbox.net\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "c22921df777de5606f1047b1345b8d22ef1c0b34", "tree": "5824c5fc27fd19086e80f080b11180ec01f1b4ea", "parents": [ "5830aa863981d43560748aa93589c0695191d95d" ], "author": { "name": "Kuniyuki Iwashima", "email": "kuniyu@amazon.com", "time": "Thu Jul 25 12:28:21 2024 -0700" }, "committer": { "name": "Pablo Neira Ayuso", "email": "pablo@netfilter.org", "time": "Wed Jul 31 23:21:34 2024 +0200" }, "message": "netfilter: iptables: Fix potential null-ptr-deref in ip6table_nat_table_init().\n\nip6table_nat_table_init() accesses net-\u003egen-\u003eptr[ip6table_nat_net_ops.id],\nbut the function is exposed to user space before the entry is allocated\nvia register_pernet_subsys().\n\nLet\u0027s call register_pernet_subsys() before xt_register_template().\n\nFixes: fdacd57c79b7 (\"netfilter: x_tables: never register tables by default\")\nSigned-off-by: Kuniyuki Iwashima \u003ckuniyu@amazon.com\u003e\nReviewed-by: Florian Westphal \u003cfw@strlen.de\u003e\nSigned-off-by: Pablo Neira Ayuso \u003cpablo@netfilter.org\u003e\n" }, { "commit": "5830aa863981d43560748aa93589c0695191d95d", "tree": "771314e701a319dbd17ee3069619662517c04d1b", "parents": [ "0bf50cead4c4710d9f704778c32ab8af47ddf070" ], "author": { "name": "Kuniyuki Iwashima", "email": "kuniyu@amazon.com", "time": "Thu Jul 25 12:28:20 2024 -0700" }, "committer": { "name": "Pablo Neira Ayuso", "email": "pablo@netfilter.org", "time": "Wed Jul 31 23:20:50 2024 +0200" }, "message": "netfilter: iptables: Fix null-ptr-deref in iptable_nat_table_init().\n\nWe had a report that iptables-restore sometimes triggered null-ptr-deref\nat boot time. [0]\n\nThe problem is that iptable_nat_table_init() is exposed to user space\nbefore the kernel fully initialises netns.\n\nIn the small race window, a user could call iptable_nat_table_init()\nthat accesses net_generic(net, iptable_nat_net_id), which is available\nonly after registering iptable_nat_net_ops.\n\nLet\u0027s call register_pernet_subsys() before xt_register_template().\n\n[0]:\nbpfilter: Loaded bpfilter_umh pid 11702\nStarted bpfilter\nBUG: kernel NULL pointer dereference, address: 0000000000000013\n PF: supervisor write access in kernel mode\n PF: error_code(0x0002) - not-present page\nPGD 0 P4D 0\nPREEMPT SMP NOPTI\nCPU: 2 PID: 11879 Comm: iptables-restor Not tainted 6.1.92-99.174.amzn2023.x86_64 #1\nHardware name: Amazon EC2 c6i.4xlarge/, BIOS 1.0 10/16/2017\nRIP: 0010:iptable_nat_table_init (net/ipv4/netfilter/iptable_nat.c:87 net/ipv4/netfilter/iptable_nat.c:121) iptable_nat\nCode: 10 4c 89 f6 48 89 ef e8 0b 19 bb ff 41 89 c4 85 c0 75 38 41 83 c7 01 49 83 c6 28 41 83 ff 04 75 dc 48 8b 44 24 08 48 8b 0c 24 \u003c48\u003e 89 08 4c 89 ef e8 a2 3b a2 cf 48 83 c4 10 44 89 e0 5b 5d 41 5c\nRSP: 0018:ffffbef902843cd0 EFLAGS: 00010246\nRAX: 0000000000000013 RBX: ffff9f4b052caa20 RCX: ffff9f4b20988d80\nRDX: 0000000000000000 RSI: 0000000000000064 RDI: ffffffffc04201c0\nRBP: ffff9f4b29394000 R08: ffff9f4b07f77258 R09: ffff9f4b07f77240\nR10: 0000000000000000 R11: ffff9f4b09635388 R12: 0000000000000000\nR13: ffff9f4b1a3c6c00 R14: ffff9f4b20988e20 R15: 0000000000000004\nFS: 00007f6284340000(0000) GS:ffff9f51fe280000(0000) knlGS:0000000000000000\nCS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033\nCR2: 0000000000000013 CR3: 00000001d10a6005 CR4: 00000000007706e0\nDR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000\nDR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400\nPKRU: 55555554\nCall Trace:\n \u003cTASK\u003e\n ? show_trace_log_lvl (arch/x86/kernel/dumpstack.c:259)\n ? show_trace_log_lvl (arch/x86/kernel/dumpstack.c:259)\n ? xt_find_table_lock (net/netfilter/x_tables.c:1259)\n ? __die_body.cold (arch/x86/kernel/dumpstack.c:478 arch/x86/kernel/dumpstack.c:420)\n ? page_fault_oops (arch/x86/mm/fault.c:727)\n ? exc_page_fault (./arch/x86/include/asm/irqflags.h:40 ./arch/x86/include/asm/irqflags.h:75 arch/x86/mm/fault.c:1470 arch/x86/mm/fault.c:1518)\n ? asm_exc_page_fault (./arch/x86/include/asm/idtentry.h:570)\n ? iptable_nat_table_init (net/ipv4/netfilter/iptable_nat.c:87 net/ipv4/netfilter/iptable_nat.c:121) iptable_nat\n xt_find_table_lock (net/netfilter/x_tables.c:1259)\n xt_request_find_table_lock (net/netfilter/x_tables.c:1287)\n get_info (net/ipv4/netfilter/ip_tables.c:965)\n ? security_capable (security/security.c:809 (discriminator 13))\n ? ns_capable (kernel/capability.c:376 kernel/capability.c:397)\n ? do_ipt_get_ctl (net/ipv4/netfilter/ip_tables.c:1656)\n ? bpfilter_send_req (net/bpfilter/bpfilter_kern.c:52) bpfilter\n nf_getsockopt (net/netfilter/nf_sockopt.c:116)\n ip_getsockopt (net/ipv4/ip_sockglue.c:1827)\n __sys_getsockopt (net/socket.c:2327)\n __x64_sys_getsockopt (net/socket.c:2342 net/socket.c:2339 net/socket.c:2339)\n do_syscall_64 (arch/x86/entry/common.c:51 arch/x86/entry/common.c:81)\n entry_SYSCALL_64_after_hwframe (arch/x86/entry/entry_64.S:121)\nRIP: 0033:0x7f62844685ee\nCode: 48 8b 0d 45 28 0f 00 f7 d8 64 89 01 48 83 c8 ff c3 66 2e 0f 1f 84 00 00 00 00 00 90 f3 0f 1e fa 49 89 ca b8 37 00 00 00 0f 05 \u003c48\u003e 3d 00 f0 ff ff 77 0a c3 66 0f 1f 84 00 00 00 00 00 48 8b 15 09\nRSP: 002b:00007ffd1f83d638 EFLAGS: 00000246 ORIG_RAX: 0000000000000037\nRAX: ffffffffffffffda RBX: 00007ffd1f83d680 RCX: 00007f62844685ee\nRDX: 0000000000000040 RSI: 0000000000000000 RDI: 0000000000000004\nRBP: 0000000000000004 R08: 00007ffd1f83d670 R09: 0000558798ffa2a0\nR10: 00007ffd1f83d680 R11: 0000000000000246 R12: 00007ffd1f83e3b2\nR13: 00007f628455baa0 R14: 00007ffd1f83d7b0 R15: 00007f628457a008\n \u003c/TASK\u003e\nModules linked in: iptable_nat(+) bpfilter rpcsec_gss_krb5 auth_rpcgss nfsv4 dns_resolver nfs lockd grace fscache veth xt_state xt_connmark xt_nat xt_statistic xt_MASQUERADE xt_mark xt_addrtype ipt_REJECT nf_reject_ipv4 nft_chain_nat nf_nat xt_conntrack nf_conntrack nf_defrag_ipv6 nf_defrag_ipv4 xt_comment nft_compat nf_tables nfnetlink overlay nls_ascii nls_cp437 vfat fat ghash_clmulni_intel aesni_intel ena crypto_simd ptp cryptd i8042 pps_core serio button sunrpc sch_fq_codel configfs loop dm_mod fuse dax dmi_sysfs crc32_pclmul crc32c_intel efivarfs\nCR2: 0000000000000013\n\nFixes: fdacd57c79b7 (\"netfilter: x_tables: never register tables by default\")\nReported-by: Takahiro Kawahara \u003ctakawaha@amazon.co.jp\u003e\nSigned-off-by: Kuniyuki Iwashima \u003ckuniyu@amazon.com\u003e\nReviewed-by: Florian Westphal \u003cfw@strlen.de\u003e\nSigned-off-by: Pablo Neira Ayuso \u003cpablo@netfilter.org\u003e\n" }, { "commit": "21b136cc63d2a9ddd60d4699552b69c214b32964", "tree": "b1fe043c07e0b7b42ad5dbc8c1127f3ded2c0ac6", "parents": [ "e4fc196f5ba36eb7b9758cf2c73df49a44199895" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jul 30 15:44:16 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Jul 31 09:57:18 2024 -0700" }, "message": "minmax: fix up min3() and max3() too\n\nDavid Laight pointed out that we should deal with the min3() and max3()\nmess too, which still does excessive expansion.\n\nAnd our current macros are actually rather broken.\n\nIn particular, the macros did this:\n\n #define min3(x, y, z) min((typeof(x))min(x, y), z)\n #define max3(x, y, z) max((typeof(x))max(x, y), z)\n\nand that not only is a nested expansion of possibly very complex\narguments with all that involves, the typing with that \"typeof()\" cast\nis completely wrong.\n\nFor example, imagine what happens in max3() if \u0027x\u0027 happens to be a\n\u0027unsigned char\u0027, but \u0027y\u0027 and \u0027z\u0027 are \u0027unsigned long\u0027. The types are\ncompatible, and there\u0027s no warning - but the result is just random\ngarbage.\n\nNo, I don\u0027t think we\u0027ve ever hit that issue in practice, but since we\nnow have sane infrastructure for doing this right, let\u0027s just use it.\nIt fixes any excessive expansion, and also avoids these kinds of broken\ntype issues.\n\nRequested-by: David Laight \u003cDavid.Laight@aculab.com\u003e\nAcked-by: Arnd Bergmann \u003carnd@kernel.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "e4fc196f5ba36eb7b9758cf2c73df49a44199895", "tree": "90cda8d1e41e2c0e34e5252976376d4f7576cbb4", "parents": [ "e254e0c5baeae28717d1b312821e6ded29e7d969", "b8e947e9f64cac9df85a07672b658df5b2bcff07" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jul 30 19:28:36 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jul 30 19:28:36 2024 -0700" }, "message": "Merge tag \u0027for-6.11-rc1-tag\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux\n\nPull btrfs fixes from David Sterba:\n\n - fix regression in extent map rework when handling insertion of\n overlapping compressed extent\n\n - fix unexpected file length when appending to a file using direct io\n and buffer not faulted in\n\n - in zoned mode, fix accounting of unusable space when flipping\n read-only block group back to read-write\n\n - fix page locking when COWing an inline range, assertion failure found\n by syzbot\n\n - fix calculation of space info in debugging print\n\n - tree-checker, add validation of data reference item\n\n - fix a few -Wmaybe-uninitialized build warnings\n\n* tag \u0027for-6.11-rc1-tag\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux:\n btrfs: initialize location to fix -Wmaybe-uninitialized in btrfs_lookup_dentry()\n btrfs: fix corruption after buffer fault in during direct IO append write\n btrfs: zoned: fix zone_unusable accounting on making block group read-write again\n btrfs: do not subtract delalloc from avail bytes\n btrfs: make cow_file_range_inline() honor locked_page on error\n btrfs: fix corrupt read due to bad offset of a compressed extent map\n btrfs: tree-checker: validate dref root and objectid\n" }, { "commit": "e254e0c5baeae28717d1b312821e6ded29e7d969", "tree": "b003c16b9236271131f4d12205103f65754ca0af", "parents": [ "c91a7dee0555f6f9d3702d86312382e4c4729d0a", "d27087c76e3c859ea05b7581ef7ce8aa5a088dd8" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jul 30 19:22:41 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jul 30 19:22:41 2024 -0700" }, "message": "Merge tag \u0027perf-tools-fixes-for-v6.11-2024-07-30\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools\n\nPull perf tools fixes from Namhyung Kim:\n \"Some more build fixes and a random crash fix:\n\n - Fix cross-build by setting pkg-config env according to the arch\n\n - Fix static build for missing library dependencies\n\n - Fix Segfault when callchain has no symbols\"\n\n* tag \u0027perf-tools-fixes-for-v6.11-2024-07-30\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools:\n perf docs: Document cross compilation\n perf: build: Link lib \u0027zstd\u0027 for static build\n perf: build: Link lib \u0027lzma\u0027 for static build\n perf: build: Only link libebl.a for old libdw\n perf: build: Set Python configuration for cross compilation\n perf: build: Setup PKG_CONFIG_LIBDIR for cross compilation\n perf tool: fix dereferencing NULL al-\u003emaps\n" }, { "commit": "0bf50cead4c4710d9f704778c32ab8af47ddf070", "tree": "dc1ccc485a0d7f77dbafb9724dc23c2ca39d51e0", "parents": [ "89add40066f9ed9abe5f7f886fe5789ff7e0c50e", "963fb4612295a5c35b1b89c8bff3bdd4f9127af6" ], "author": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Tue Jul 30 18:41:10 2024 -0700" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Tue Jul 30 18:41:10 2024 -0700" }, "message": "Merge branch \u0027100GbE\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net-queue\n\nTony Nguyen says:\n\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\nice: fix AF_XDP ZC timeout and concurrency issues\n\nMaciej Fijalkowski says:\n\nChanges included in this patchset address an issue that customer has\nbeen facing when AF_XDP ZC Tx sockets were used in combination with flow\ncontrol and regular Tx traffic.\n\nAfter executing:\nethtool --set-priv-flags $dev link-down-on-close on\nethtool -A $dev rx on tx on\n\nlaunching multiple ZC Tx sockets on $dev + pinging remote interface (so\nthat regular Tx traffic is present) and then going through down/up of\n$dev, Tx timeout occurred and then most of the time ice driver was unable\nto recover from that state.\n\nThese patches combined together solve the described above issue on\ncustomer side. Main focus here is to forbid producing Tx descriptors when\neither carrier is not yet initialized or process of bringing interface\ndown has already started.\n\nv1: https://lore.kernel.org/netdev/20240708221416.625850-1-anthony.l.nguyen@intel.com/\n\n* \u0027100GbE\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net-queue:\n ice: xsk: fix txq interrupt mapping\n ice: add missing WRITE_ONCE when clearing ice_rx_ring::xdp_prog\n ice: improve updating ice_{t,r}x_ring::xsk_pool\n ice: toggle netif_carrier when setting up XSK pool\n ice: modify error handling when setting XSK pool in ndo_bpf\n ice: replace synchronize_rcu with synchronize_net\n ice: don\u0027t busy wait for Rx queue disable in ice_qp_dis()\n ice: respect netif readiness in AF_XDP ZC related ndo\u0027s\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\nLink: https://patch.msgid.link/20240729200716.681496-1-anthony.l.nguyen@intel.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "89add40066f9ed9abe5f7f886fe5789ff7e0c50e", "tree": "0480a24ff9eaf11efbf8ee66ef07eda2716666a4", "parents": [ "a7f3abcf635767b2e19c4c55c4c35756595ebc86" ], "author": { "name": "Willem de Bruijn", "email": "willemb@google.com", "time": "Mon Jul 29 16:10:12 2024 -0400" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Tue Jul 30 18:34:13 2024 -0700" }, "message": "net: drop bad gso csum_start and offset in virtio_net_hdr\n\nTighten csum_start and csum_offset checks in virtio_net_hdr_to_skb\nfor GSO packets.\n\nThe function already checks that a checksum requested with\nVIRTIO_NET_HDR_F_NEEDS_CSUM is in skb linear. But for GSO packets\nthis might not hold for segs after segmentation.\n\nSyzkaller demonstrated to reach this warning in skb_checksum_help\n\n\toffset \u003d skb_checksum_start_offset(skb);\n\tret \u003d -EINVAL;\n\tif (WARN_ON_ONCE(offset \u003e\u003d skb_headlen(skb)))\n\nBy injecting a TSO packet:\n\nWARNING: CPU: 1 PID: 3539 at net/core/dev.c:3284 skb_checksum_help+0x3d0/0x5b0\n ip_do_fragment+0x209/0x1b20 net/ipv4/ip_output.c:774\n ip_finish_output_gso net/ipv4/ip_output.c:279 [inline]\n __ip_finish_output+0x2bd/0x4b0 net/ipv4/ip_output.c:301\n iptunnel_xmit+0x50c/0x930 net/ipv4/ip_tunnel_core.c:82\n ip_tunnel_xmit+0x2296/0x2c70 net/ipv4/ip_tunnel.c:813\n __gre_xmit net/ipv4/ip_gre.c:469 [inline]\n ipgre_xmit+0x759/0xa60 net/ipv4/ip_gre.c:661\n __netdev_start_xmit include/linux/netdevice.h:4850 [inline]\n netdev_start_xmit include/linux/netdevice.h:4864 [inline]\n xmit_one net/core/dev.c:3595 [inline]\n dev_hard_start_xmit+0x261/0x8c0 net/core/dev.c:3611\n __dev_queue_xmit+0x1b97/0x3c90 net/core/dev.c:4261\n packet_snd net/packet/af_packet.c:3073 [inline]\n\nThe geometry of the bad input packet at tcp_gso_segment:\n\n[ 52.003050][ T8403] skb len\u003d12202 headroom\u003d244 headlen\u003d12093 tailroom\u003d0\n[ 52.003050][ T8403] mac\u003d(168,24) mac_len\u003d24 net\u003d(192,52) trans\u003d244\n[ 52.003050][ T8403] shinfo(txflags\u003d0 nr_frags\u003d1 gso(size\u003d1552 type\u003d3 segs\u003d0))\n[ 52.003050][ T8403] csum(0x60000c7 start\u003d199 offset\u003d1536\nip_summed\u003d3 complete_sw\u003d0 valid\u003d0 level\u003d0)\n\nMitigate with stricter input validation.\n\ncsum_offset: for GSO packets, deduce the correct value from gso_type.\nThis is already done for USO. Extend it to TSO. Let UFO be:\nudp[46]_ufo_fragment ignores these fields and always computes the\nchecksum in software.\n\ncsum_start: finding the real offset requires parsing to the transport\nheader. Do not add a parser, use existing segmentation parsing. Thanks\nto SKB_GSO_DODGY, that also catches bad packets that are hw offloaded.\nAgain test both TSO and USO. Do not test UFO for the above reason, and\ndo not test UDP tunnel offload.\n\nGSO packet are almost always CHECKSUM_PARTIAL. USO packets may be\nCHECKSUM_NONE since commit 10154dbded6d6 (\"udp: Allow GSO transmit\nfrom devices with no checksum offload\"), but then still these fields\nare initialized correctly in udp4_hwcsum/udp6_hwcsum_outgoing. So no\nneed to test for ip_summed \u003d\u003d CHECKSUM_PARTIAL first.\n\nThis revises an existing fix mentioned in the Fixes tag, which broke\nsmall packets with GSO offload, as detected by kselftests.\n\nLink: https://syzkaller.appspot.com/bug?extid\u003de1db31216c789f552871\nLink: https://lore.kernel.org/netdev/20240723223109.2196886-1-kuba@kernel.org\nFixes: e269d79c7d35 (\"net: missing check virtio\")\nCc: stable@vger.kernel.org\nSigned-off-by: Willem de Bruijn \u003cwillemb@google.com\u003e\nLink: https://patch.msgid.link/20240729201108.1615114-1-willemdebruijn.kernel@gmail.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "a7f3abcf635767b2e19c4c55c4c35756595ebc86", "tree": "9d1527146f6c7eb823a4616c5c87b506cc48fa32", "parents": [ "84383b5ef4cd21b4a67de92afdc05a03b5247db9" ], "author": { "name": "Bartosz Golaszewski", "email": "bartosz.golaszewski@linaro.org", "time": "Mon Jul 29 17:03:14 2024 +0200" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Tue Jul 30 18:27:58 2024 -0700" }, "message": "net: phy: aquantia: only poll GLOBAL_CFG regs on aqr113, aqr113c and aqr115c\n\nCommit 708405f3e56e (\"net: phy: aquantia: wait for the GLOBAL_CFG to\nstart returning real values\") introduced a workaround for an issue\nobserved on aqr115c. However there were never any reports of it\nhappening on other models and the workaround has been reported to cause\nand issue on aqr113c (and it may cause the same on any other model not\nsupporting 10M mode).\n\nLet\u0027s limit the impact of the workaround to aqr113, aqr113c and aqr115c\nand poll the 100M GLOBAL_CFG register instead as both models are known\nto support it correctly.\n\nReported-by: Jon Hunter \u003cjonathanh@nvidia.com\u003e\nCloses: https://lore.kernel.org/lkml/7c0140be-4325-4005-9068-7e0fc5ff344d@nvidia.com/\nFixes: 708405f3e56e (\"net: phy: aquantia: wait for the GLOBAL_CFG to start returning real values\")\nTested-by: Jon Hunter \u003cjonathanh@nvidia.com\u003e\nSigned-off-by: Bartosz Golaszewski \u003cbartosz.golaszewski@linaro.org\u003e\nReviewed-by: Antoine Tenart \u003catenart@kernel.org\u003e\nLink: https://patch.msgid.link/20240729150315.65798-1-brgl@bgdev.pl\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "84383b5ef4cd21b4a67de92afdc05a03b5247db9", "tree": "0c7c45d07ead5424596da28e7a76712cc5b7bac1", "parents": [ "0aa3ca956c46d849775eae1816cef8fe4bc8b50e" ], "author": { "name": "Raju Lakkaraju", "email": "Raju.Lakkaraju@microchip.com", "time": "Thu Jul 25 12:41:25 2024 +0530" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Tue Jul 30 15:10:53 2024 -0700" }, "message": "net: phy: micrel: Fix the KSZ9131 MDI-X status issue\n\nThe MDIX status is not accurately reflecting the current state after the link\npartner has manually altered its MDIX configuration while operating in forced\nmode.\n\nAccess information about Auto mdix completion and pair selection from the\nKSZ9131\u0027s Auto/MDI/MDI-X status register\n\nFixes: b64e6a8794d9 (\"net: phy: micrel: Add PHY Auto/MDI/MDI-X set driver for KSZ9131\")\nSigned-off-by: Raju Lakkaraju \u003cRaju.Lakkaraju@microchip.com\u003e\nReviewed-by: Andrew Lunn \u003candrew@lunn.ch\u003e\nLink: https://patch.msgid.link/20240725071125.13960-1-Raju.Lakkaraju@microchip.com\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "7764b9622db4382b2797b54a70f292c8da6ef417", "tree": "849b01b4dc7b6094ed79beb752130f697dc24088", "parents": [ "41c24102af7b6236277a214428b203d51a3462df" ], "author": { "name": "Jiri Olsa", "email": "jolsa@kernel.org", "time": "Fri Jul 26 20:08:47 2024 +0200" }, "committer": { "name": "Andrii Nakryiko", "email": "andrii@kernel.org", "time": "Tue Jul 30 13:42:24 2024 -0700" }, "message": "bpf/selftests: Fix ASSERT_OK condition check in uprobe_syscall test\n\nFixing ASSERT_OK condition check in uprobe_syscall test,\notherwise we return from test on pipe success.\n\nSigned-off-by: Jiri Olsa \u003cjolsa@kernel.org\u003e\nSigned-off-by: Andrii Nakryiko \u003candrii@kernel.org\u003e\nAcked-by: Yonghong Song \u003cyonghong.song@linux.dev\u003e\nLink: https://lore.kernel.org/bpf/20240726180847.684584-1-jolsa@kernel.org\n" }, { "commit": "c91a7dee0555f6f9d3702d86312382e4c4729d0a", "tree": "eb2663a5e93191c73fbf00a8ca6b811b41b63720", "parents": [ "22f5468731491e53356ba7c028f0fdea20b18e2c", "df615907f1bf907260af01ccb904d0e9304b5278" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jul 30 12:53:52 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jul 30 12:53:52 2024 -0700" }, "message": "Merge tag \u0027chrome-platform-fixes-for-v6.11-rc2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/chrome-platform/linux\n\nPull chrome-platform fix from Tzung-Bi Shih:\n \"Fix a race condition that sends multiple host commands at a time\"\n\n* tag \u0027chrome-platform-fixes-for-v6.11-rc2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/chrome-platform/linux:\n platform/chrome: cros_ec_proto: Lock device when updating MKBP version\n" }, { "commit": "22f5468731491e53356ba7c028f0fdea20b18e2c", "tree": "510971ee62c41a8f1d9a2fe4e51f5f3641198455", "parents": [ "94ede2a3e9135764736221c080ac7c0ad993dc2d" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jul 30 10:36:47 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jul 30 10:36:47 2024 -0700" }, "message": "minmax: improve macro expansion and type checking\n\nThis clarifies the rules for min()/max()/clamp() type checking and makes\nthem a much more efficient macro expansion.\n\nIn particular, we now look at the type and range of the inputs to see\nwhether they work together, generating a mask of acceptable comparisons,\nand then just verifying that the inputs have a shared case:\n\n - an expression with a signed type can be used for\n (1) signed comparisons\n (2) unsigned comparisons if it is statically known to have a\n non-negative value\n\n - an expression with an unsigned type can be used for\n (3) unsigned comparison\n (4) signed comparisons if the type is smaller than \u0027int\u0027 and thus\n the C integer promotion rules will make it signed anyway\n\nHere rule (1) and (3) are obvious, and rule (2) is important in order to\nallow obvious trivial constants to be used together with unsigned\nvalues.\n\nRule (4) is not necessarily a good idea, but matches what we used to do,\nand we have extant cases of this situation in the kernel. Notably with\nbcachefs having an expression like\n\n\tmin(bch2_bucket_sectors_dirty(a), ca-\u003emi.bucket_size)\n\nwhere bch2_bucket_sectors_dirty() returns an \u0027s64\u0027, and\n\u0027ca-\u003emi.bucket_size\u0027 is of type \u0027u16\u0027.\n\nTechnically that bcachefs comparison is clearly sensible on a C type\nlevel, because the \u0027u16\u0027 will go through the normal C integer promotion,\nand become \u0027int\u0027, and then we\u0027re comparing two signed values and\neverything looks sane.\n\nHowever, it\u0027s not entirely clear that a \u0027min(s64,u16)\u0027 operation makes a\nlot of conceptual sense, and it\u0027s possible that we will remove rule (4).\nAfter all, the _reason_ we have these complicated type checks is exactly\nthat the C type promotion rules are not very intuitive.\n\nBut at least for now the rule is in place for backwards compatibility.\n\nAlso note that rule (2) existed before, but is hugely relaxed by this\ncommit. It used to be true only for the simplest compile-time\nnon-negative integer constants. The new macro model will allow cases\nwhere the compiler can trivially see that an expression is non-negative\neven if it isn\u0027t necessarily a constant.\n\nFor example, the amdgpu driver does\n\n\tmin_t(size_t, sizeof(fru_info-\u003eserial), pia[addr] \u0026 0x3F));\n\nbecause our old \u0027min()\u0027 macro would see that \u0027pia[addr] \u0026 0x3F\u0027 is of\ntype \u0027int\u0027 and clearly not a C constant expression, so doing a \u0027min()\u0027\nwith a \u0027size_t\u0027 is a signedness violation.\n\nOur new \u0027min()\u0027 macro still sees that \u0027pia[addr] \u0026 0x3F\u0027 is of type\n\u0027int\u0027, but is smart enough to also see that it is clearly non-negative,\nand thus would allow that case without any complaints.\n\nCc: Arnd Bergmann \u003carnd@kernel.org\u003e\nCc: David Laight \u003cDavid.Laight@aculab.com\u003e\nCc: Lorenzo Stoakes \u003clorenzo.stoakes@oracle.com\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "0aa3ca956c46d849775eae1816cef8fe4bc8b50e", "tree": "da6f99e7e7fa210545e2610dfa0dda6f463fd44f", "parents": [ "f558120cd709682b739207b48cf7479fd9568431" ], "author": { "name": "Dan Carpenter", "email": "dan.carpenter@linaro.org", "time": "Wed Jul 24 11:06:56 2024 -0500" }, "committer": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Tue Jul 30 09:00:12 2024 -0700" }, "message": "net: mvpp2: Don\u0027t re-use loop iterator\n\nThis function has a nested loop. The problem is that both the inside\nand outside loop use the same variable as an iterator. I found this\nvia static analysis so I\u0027m not sure the impact. It could be that it\nloops forever or, more likely, the loop exits early.\n\nFixes: 3a616b92a9d1 (\"net: mvpp2: Add TX flow control support for jumbo frames\")\nSigned-off-by: Dan Carpenter \u003cdan.carpenter@linaro.org\u003e\nReviewed-by: Simon Horman \u003chorms@kernel.org\u003e\nLink: https://patch.msgid.link/eaa8f403-7779-4d81-973d-a9ecddc0bf6f@stanley.mountain\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\n" }, { "commit": "b8e947e9f64cac9df85a07672b658df5b2bcff07", "tree": "3ced7284c4ee5cd3ab531a502c9949e7f7a5d40b", "parents": [ "939b656bc8ab203fdbde26ccac22bcb7f0985be5" ], "author": { "name": "David Sterba", "email": "dsterba@suse.com", "time": "Mon Jul 29 21:59:24 2024 +0200" }, "committer": { "name": "David Sterba", "email": "dsterba@suse.com", "time": "Tue Jul 30 15:33:06 2024 +0200" }, "message": "btrfs: initialize location to fix -Wmaybe-uninitialized in btrfs_lookup_dentry()\n\nSome arch + compiler combinations report a potentially unused variable\nlocation in btrfs_lookup_dentry(). This is a false alert as the variable\nis passed by value and always valid or there\u0027s an error. The compilers\ncannot probably reason about that although btrfs_inode_by_name() is in\nthe same file.\n\n \u003e + /kisskb/src/fs/btrfs/inode.c: error: \u0027location.objectid\u0027 may be used\n +uninitialized in this function [-Werror\u003dmaybe-uninitialized]: \u003d\u003e 5603:9\n \u003e + /kisskb/src/fs/btrfs/inode.c: error: \u0027location.type\u0027 may be used\n +uninitialized in this function [-Werror\u003dmaybe-uninitialized]: \u003d\u003e 5674:5\n\n m68k-gcc8/m68k-allmodconfig\n mips-gcc8/mips-allmodconfig\n powerpc-gcc5/powerpc-all{mod,yes}config\n powerpc-gcc5/ppc64_defconfig\n\nInitialize it to zero, this should fix the warnings and won\u0027t change the\nbehaviour as btrfs_inode_by_name() accepts only a root or inode item\ntypes, otherwise returns an error.\n\nReported-by: Geert Uytterhoeven \u003cgeert@linux-m68k.org\u003e\nTested-by: Geert Uytterhoeven \u003cgeert@linux-m68k.org\u003e\nLink: https://lore.kernel.org/linux-btrfs/bd4e9928-17b3-9257-8ba7-6b7f9bbb639a@linux-m68k.org/\nReviewed-by: Qu Wenruo \u003cwqu@suse.com\u003e\nSigned-off-by: David Sterba \u003cdsterba@suse.com\u003e\n" }, { "commit": "f558120cd709682b739207b48cf7479fd9568431", "tree": "6c5d9a491a79946c4fd64975207c94c55631f75d", "parents": [ "2fe5273f149cc882c371f9954b5fdbd1bd8c9b5c" ], "author": { "name": "Alexandra Winter", "email": "wintera@linux.ibm.com", "time": "Mon Jul 29 14:28:16 2024 +0200" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Tue Jul 30 15:01:50 2024 +0200" }, "message": "net/iucv: fix use after free in iucv_sock_close()\n\niucv_sever_path() is called from process context and from bh context.\niucv-\u003epath is used as indicator whether somebody else is taking care of\nsevering the path (or it is already removed / never existed).\nThis needs to be done with atomic compare and swap, otherwise there is a\nsmall window where iucv_sock_close() will try to work with a path that has\nalready been severed and freed by iucv_callback_connrej() called by\niucv_tasklet_fn().\n\nExample:\n[452744.123844] Call Trace:\n[452744.123845] ([\u003c0000001e87f03880\u003e] 0x1e87f03880)\n[452744.123966] [\u003c00000000d593001e\u003e] iucv_path_sever+0x96/0x138\n[452744.124330] [\u003c000003ff801ddbca\u003e] iucv_sever_path+0xc2/0xd0 [af_iucv]\n[452744.124336] [\u003c000003ff801e01b6\u003e] iucv_sock_close+0xa6/0x310 [af_iucv]\n[452744.124341] [\u003c000003ff801e08cc\u003e] iucv_sock_release+0x3c/0xd0 [af_iucv]\n[452744.124345] [\u003c00000000d574794e\u003e] __sock_release+0x5e/0xe8\n[452744.124815] [\u003c00000000d5747a0c\u003e] sock_close+0x34/0x48\n[452744.124820] [\u003c00000000d5421642\u003e] __fput+0xba/0x268\n[452744.124826] [\u003c00000000d51b382c\u003e] task_work_run+0xbc/0xf0\n[452744.124832] [\u003c00000000d5145710\u003e] do_notify_resume+0x88/0x90\n[452744.124841] [\u003c00000000d5978096\u003e] system_call+0xe2/0x2c8\n[452744.125319] Last Breaking-Event-Address:\n[452744.125321] [\u003c00000000d5930018\u003e] iucv_path_sever+0x90/0x138\n[452744.125324]\n[452744.125325] Kernel panic - not syncing: Fatal exception in interrupt\n\nNote that bh_lock_sock() is not serializing the tasklet context against\nprocess context, because the check for sock_owned_by_user() and\ncorresponding handling is missing.\n\nIdeas for a future clean-up patch:\nA) Correct usage of bh_lock_sock() in tasklet context, as described in\nLink: https://lore.kernel.org/netdev/1280155406.2899.407.camel@edumazet-laptop/\nRe-enqueue, if needed. This may require adding return values to the\ntasklet functions and thus changes to all users of iucv.\n\nB) Change iucv tasklet into worker and use only lock_sock() in af_iucv.\n\nFixes: 7d316b945352 (\"af_iucv: remove IUCV-pathes completely\")\nReviewed-by: Halil Pasic \u003cpasic@linux.ibm.com\u003e\nSigned-off-by: Alexandra Winter \u003cwintera@linux.ibm.com\u003e\nLink: https://patch.msgid.link/20240729122818.947756-1-wintera@linux.ibm.com\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n\n" }, { "commit": "df615907f1bf907260af01ccb904d0e9304b5278", "tree": "65004f50a72021a5cc46f554f14a46463117e844", "parents": [ "4baf1cc54433ff7c6e5178517bc8768001416681" ], "author": { "name": "Patryk Duda", "email": "patrykd@google.com", "time": "Tue Jul 30 10:44:25 2024 +0000" }, "committer": { "name": "Tzung-Bi Shih", "email": "tzungbi@kernel.org", "time": "Tue Jul 30 19:48:35 2024 +0800" }, "message": "platform/chrome: cros_ec_proto: Lock device when updating MKBP version\n\nThe cros_ec_get_host_command_version_mask() function requires that the\ncaller must have ec_dev-\u003elock mutex before calling it. This requirement\nwas not met and as a result it was possible that two commands were sent\nto the device at the same time.\n\nThe problem was observed while using UART backend which doesn\u0027t use any\nadditional locks, unlike SPI backend which locks the controller until\nresponse is received.\n\nFixes: f74c7557ed0d (\"platform/chrome: cros_ec_proto: Update version on GET_NEXT_EVENT failure\")\nCc: stable@vger.kernel.org\nSigned-off-by: Patryk Duda \u003cpatrykd@google.com\u003e\nLink: https://lore.kernel.org/r/20240730104425.607083-1-patrykd@google.com\nSigned-off-by: Tzung-Bi Shih \u003ctzungbi@kernel.org\u003e\n" }, { "commit": "2fe5273f149cc882c371f9954b5fdbd1bd8c9b5c", "tree": "8147803cd8b4d42f4dbac4c97bb917dbcee00b59", "parents": [ "0cd55ef92a4a94fabc6e7d5fb2e8dfced1860d88" ], "author": { "name": "D. Wythe", "email": "alibuda@linux.alibaba.com", "time": "Mon Jul 29 11:40:15 2024 +0800" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Tue Jul 30 11:12:25 2024 +0200" }, "message": "net/smc: prevent UAF in inet_create()\n\nFollowing syzbot repro crashes the kernel:\n\nsocketpair(0x2, 0x1, 0x100, \u0026(0x7f0000000140)) (fail_nth: 13)\n\nFix this by not calling sk_common_release() from smc_create_clcsk().\n\nStack trace:\nsocket: no more sockets\n------------[ cut here ]------------\nrefcount_t: underflow; use-after-free.\n WARNING: CPU: 1 PID: 5092 at lib/refcount.c:28\nrefcount_warn_saturate+0x15a/0x1d0 lib/refcount.c:28\nModules linked in:\nCPU: 1 PID: 5092 Comm: syz-executor424 Not tainted\n6.10.0-syzkaller-04483-g0be9ae5486cd #0\nHardware name: Google Google Compute Engine/Google Compute Engine, BIOS\nGoogle 06/27/2024\n RIP: 0010:refcount_warn_saturate+0x15a/0x1d0 lib/refcount.c:28\nCode: 80 f3 1f 8c e8 e7 69 a8 fc 90 0f 0b 90 90 eb 99 e8 cb 4f e6 fc c6\n05 8a 8d e8 0a 01 90 48 c7 c7 e0 f3 1f 8c e8 c7 69 a8 fc 90 \u003c0f\u003e 0b 90\n90 e9 76 ff ff ff e8 a8 4f e6 fc c6 05 64 8d e8 0a 01 90\nRSP: 0018:ffffc900034cfcf0 EFLAGS: 00010246\nRAX: 3b9fcde1c862f700 RBX: ffff888022918b80 RCX: ffff88807b39bc00\nRDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000000\nRBP: 0000000000000003 R08: ffffffff815878a2 R09: fffffbfff1c39d94\nR10: dffffc0000000000 R11: fffffbfff1c39d94 R12: 00000000ffffffe9\nR13: 1ffff11004523165 R14: ffff888022918b28 R15: ffff888022918b00\nFS: 00005555870e7380(0000) GS:ffff8880b9500000(0000)\nknlGS:0000000000000000\nCS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033\nCR2: 0000000020000140 CR3: 000000007582e000 CR4: 00000000003506f0\nDR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000\nDR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400\nCall Trace:\n \u003cTASK\u003e\n inet_create+0xbaf/0xe70\n __sock_create+0x490/0x920 net/socket.c:1571\n sock_create net/socket.c:1622 [inline]\n __sys_socketpair+0x2ca/0x720 net/socket.c:1769\n __do_sys_socketpair net/socket.c:1822 [inline]\n __se_sys_socketpair net/socket.c:1819 [inline]\n __x64_sys_socketpair+0x9b/0xb0 net/socket.c:1819\n do_syscall_x64 arch/x86/entry/common.c:52 [inline]\n do_syscall_64+0xf3/0x230 arch/x86/entry/common.c:83\n entry_SYSCALL_64_after_hwframe+0x77/0x7f\nRIP: 0033:0x7fbcb9259669\nCode: 28 00 00 00 75 05 48 83 c4 28 c3 e8 a1 1a 00 00 90 48 89 f8 48 89\nf7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 \u003c48\u003e 3d 01\nf0 ff ff 73 01 c3 48 c7 c1 b8 ff ff ff f7 d8 64 89 01 48\nRSP: 002b:00007fffe931c6d8 EFLAGS: 00000246 ORIG_RAX: 0000000000000035\nRAX: ffffffffffffffda RBX: 00007fffe931c6f0 RCX: 00007fbcb9259669\nRDX: 0000000000000100 RSI: 0000000000000001 RDI: 0000000000000002\nRBP: 0000000000000002 R08: 00007fffe931c476 R09: 00000000000000a0\nR10: 0000000020000140 R11: 0000000000000246 R12: 00007fffe931c6ec\nR13: 431bde82d7b634db R14: 0000000000000001 R15: 0000000000000001\n \u003c/TASK\u003e\n\nLink: https://lore.kernel.org/r/20240723175809.537291-1-edumazet@google.com/\nFixes: d25a92ccae6b (\"net/smc: Introduce IPPROTO_SMC\")\nReported-by: syzbot \u003csyzkaller@googlegroups.com\u003e\nSigned-off-by: D. Wythe \u003calibuda@linux.alibaba.com\u003e\nReviewed-by: Eric Dumazet \u003cedumazet@google.com\u003e\nReviewed-by: Wenjia Zhang \u003cwenjia@linux.ibm.com\u003e\nLink: https://patch.msgid.link/1722224415-30999-1-git-send-email-alibuda@linux.alibaba.com\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n\n" }, { "commit": "0cd55ef92a4a94fabc6e7d5fb2e8dfced1860d88", "tree": "38872912cb36c48fb728d3a04d114686af1a9998", "parents": [ "039564d2fd37b122ec0d268e2ee6334e7169e225", "f833470c27832136d4416d8fc55d658082af0989" ], "author": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Tue Jul 30 10:27:32 2024 +0200" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Tue Jul 30 10:28:13 2024 +0200" }, "message": "Merge branch \u0027mptcp-fix-inconsistent-backup-usage\u0027\n\nMatthieu Baerts says:\n\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\nmptcp: fix inconsistent backup usage\n\nIn all the MPTCP backup related tests, the backup flag was set on one\nside, and the expected behaviour is to have both sides respecting this\ndecision. That\u0027s also the \"natural\" way, and what the users seem to\nexpect.\n\nOn the scheduler side, only the \u0027backup\u0027 field was checked, which is\nsupposed to be set only if the other peer flagged a subflow as backup.\nBut in various places, this flag was also set when the local host\nflagged the subflow as backup, certainly to have the expected behaviour\nmentioned above.\n\nPatch 1 modifies the packet scheduler to check if the backup flag has\nbeen set on both directions, not to change its behaviour after having\napplied the following patches. That\u0027s what the default packet scheduler\nshould have done since the beginning in v5.7.\n\nPatch 2 fixes the backup flag being mirrored on the MPJ+SYN+ACK by\naccident since its introduction in v5.7. Instead, the received and sent\nbackup flags are properly distinguished in requests.\n\nPatch 3 stops setting the received backup flag as well when sending an\nMP_PRIO, something that was done since the MP_PRIO support in v5.12.\n\nPatch 4 adds related and missing MIB counters to be able to easily check\nif MP_JOIN are sent with a backup flag. Certainly because these counters\nwere not there, the behaviour that is fixed by patches here was not\nproperly verified.\n\nPatch 5 validates the previous patch by extending the MPTCP Join\nselftest.\n\nPatch 6 fixes the backup support in signal endpoints: if a signal\nendpoint had the backup flag, it was not set in the MPJ+SYN+ACK as\nexpected. It was only set for ongoing connections, but not future ones\nas expected, since the introduction of the backup flag in endpoints in\nv5.10.\n\nPatch 7 validates the previous patch by extending the MPTCP Join\nselftest as well.\n\nSigned-off-by: Matthieu Baerts (NGI0) \u003cmatttbe@kernel.org\u003e\n---\nMatthieu Baerts (NGI0) (7):\n mptcp: sched: check both directions for backup\n mptcp: distinguish rcv vs sent backup flag in requests\n mptcp: pm: only set request_bkup flag when sending MP_PRIO\n mptcp: mib: count MPJ with backup flag\n selftests: mptcp: join: validate backup in MPJ\n mptcp: pm: fix backup support in signal endpoints\n selftests: mptcp: join: check backup support in signal endp\n\n include/trace/events/mptcp.h | 2 +-\n net/mptcp/mib.c | 2 +\n net/mptcp/mib.h | 2 +\n net/mptcp/options.c | 2 +-\n net/mptcp/pm.c | 12 +++++\n net/mptcp/pm_netlink.c | 19 ++++++-\n net/mptcp/pm_userspace.c | 18 +++++++\n net/mptcp/protocol.c | 10 ++--\n net/mptcp/protocol.h | 4 ++\n net/mptcp/subflow.c | 10 ++++\n tools/testing/selftests/net/mptcp/mptcp_join.sh | 72 ++++++++++++++++++++-----\n 11 files changed, 132 insertions(+), 21 deletions(-)\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\nLink: https://patch.msgid.link/20240727-upstream-net-20240727-mptcp-backup-signal-v1-0-f50b31604cf1@kernel.org\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n" }, { "commit": "f833470c27832136d4416d8fc55d658082af0989", "tree": "38872912cb36c48fb728d3a04d114686af1a9998", "parents": [ "6834097fc38c5416701c793da94558cea49c0a1f" ], "author": { "name": "Matthieu Baerts (NGI0)", "email": "matttbe@kernel.org", "time": "Sat Jul 27 12:01:29 2024 +0200" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Tue Jul 30 10:27:30 2024 +0200" }, "message": "selftests: mptcp: join: check backup support in signal endp\n\nBefore the previous commit, \u0027signal\u0027 endpoints with the \u0027backup\u0027 flag\nwere ignored when sending the MP_JOIN.\n\nThe MPTCP Join selftest has then been modified to validate this case:\nthe \"single address, backup\" test, is now validating the MP_JOIN with a\nbackup flag as it is what we expect it to do with such name. The\nprevious version has been kept, but renamed to \"single address, switch\nto backup\" to avoid confusions.\n\nThe \"single address with port, backup\" test is also now validating the\nMPJ with a backup flag, which makes more sense than checking the switch\nto backup with an MP_PRIO.\n\nThe \"mpc backup both sides\" test is now validating that the backup flag\nis also set in MP_JOIN from and to the addresses used in the initial\nsubflow, using the special ID 0.\n\nThe \u0027Fixes\u0027 tag here below is the same as the one from the previous\ncommit: this patch here is not fixing anything wrong in the selftests,\nbut it validates the previous fix for an issue introduced by this commit\nID.\n\nFixes: 4596a2c1b7f5 (\"mptcp: allow creating non-backup subflows\")\nCc: stable@vger.kernel.org\nReviewed-by: Mat Martineau \u003cmartineau@kernel.org\u003e\nSigned-off-by: Matthieu Baerts (NGI0) \u003cmatttbe@kernel.org\u003e\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n\n" }, { "commit": "6834097fc38c5416701c793da94558cea49c0a1f", "tree": "819f07d69be352b553b2ad9866e3d53abd2bc689", "parents": [ "935ff5bb8a1cfcdf8e60c8f5c794d0bbbc234437" ], "author": { "name": "Matthieu Baerts (NGI0)", "email": "matttbe@kernel.org", "time": "Sat Jul 27 12:01:28 2024 +0200" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Tue Jul 30 10:27:30 2024 +0200" }, "message": "mptcp: pm: fix backup support in signal endpoints\n\nThere was a support for signal endpoints, but only when the endpoint\u0027s\nflag was changed during a connection. If an endpoint with the signal and\nbackup was already present, the MP_JOIN reply was not containing the\nbackup flag as expected.\n\nThat\u0027s confusing to have this inconsistent behaviour. On the other hand,\nthe infrastructure to set the backup flag in the SYN + ACK + MP_JOIN was\nalready there, it was just never set before. Now when requesting the\nlocal ID from the path-manager, the backup status is also requested.\n\nNote that when the userspace PM is used, the backup flag can be set if\nthe local address was already used before with a backup flag, e.g. if\nthe address was announced with the \u0027backup\u0027 flag, or a subflow was\ncreated with the \u0027backup\u0027 flag.\n\nFixes: 4596a2c1b7f5 (\"mptcp: allow creating non-backup subflows\")\nCc: stable@vger.kernel.org\nCloses: https://github.com/multipath-tcp/mptcp_net-next/issues/507\nReviewed-by: Mat Martineau \u003cmartineau@kernel.org\u003e\nSigned-off-by: Matthieu Baerts (NGI0) \u003cmatttbe@kernel.org\u003e\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n\n" }, { "commit": "935ff5bb8a1cfcdf8e60c8f5c794d0bbbc234437", "tree": "3f84b46a85fe0540277061fe3bc913fe7f2dd4ca", "parents": [ "4dde0d72ccec500c60c798e036b852e013d6e124" ], "author": { "name": "Matthieu Baerts (NGI0)", "email": "matttbe@kernel.org", "time": "Sat Jul 27 12:01:27 2024 +0200" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Tue Jul 30 10:27:30 2024 +0200" }, "message": "selftests: mptcp: join: validate backup in MPJ\n\nA peer can notify the other one that a subflow has to be treated as\n\"backup\" by two different ways: either by sending a dedicated MP_PRIO\nnotification, or by setting the backup flag in the MP_JOIN handshake.\n\nThe selftests were previously monitoring the former, but not the latter.\nThis is what is now done here by looking at these new MIB counters when\nvalidating the \u0027backup\u0027 cases:\n\n MPTcpExtMPJoinSynBackupRx\n MPTcpExtMPJoinSynAckBackupRx\n\nThe \u0027Fixes\u0027 tag here below is the same as the one from the previous\ncommit: this patch here is not fixing anything wrong in the selftests,\nbut it will help to validate a new fix for an issue introduced by this\ncommit ID.\n\nFixes: 4596a2c1b7f5 (\"mptcp: allow creating non-backup subflows\")\nCc: stable@vger.kernel.org\nReviewed-by: Mat Martineau \u003cmartineau@kernel.org\u003e\nSigned-off-by: Matthieu Baerts (NGI0) \u003cmatttbe@kernel.org\u003e\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n\n" }, { "commit": "4dde0d72ccec500c60c798e036b852e013d6e124", "tree": "e4f1b9a40bc4f0f2aac8acfa30f8fb0a31509903", "parents": [ "4258b94831bb7ff28ab80e3c8d94db37db930728" ], "author": { "name": "Matthieu Baerts (NGI0)", "email": "matttbe@kernel.org", "time": "Sat Jul 27 12:01:26 2024 +0200" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Tue Jul 30 10:27:30 2024 +0200" }, "message": "mptcp: mib: count MPJ with backup flag\n\nWithout such counters, it is difficult to easily debug issues with MPJ\nnot having the backup flags on production servers.\n\nThis is not strictly a fix, but it eases to validate the following\npatches without requiring to take packet traces, to query ongoing\nconnections with Netlink with admin permissions, or to guess by looking\nat the behaviour of the packet scheduler. Also, the modification is self\ncontained, isolated, well controlled, and the increments are done just\nafter others, there from the beginning. It looks then safe, and helpful\nto backport this.\n\nFixes: 4596a2c1b7f5 (\"mptcp: allow creating non-backup subflows\")\nCc: stable@vger.kernel.org\nReviewed-by: Mat Martineau \u003cmartineau@kernel.org\u003e\nSigned-off-by: Matthieu Baerts (NGI0) \u003cmatttbe@kernel.org\u003e\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n\n" }, { "commit": "4258b94831bb7ff28ab80e3c8d94db37db930728", "tree": "d50cff4945d3641484b9b46d589ad93f440a4471", "parents": [ "efd340bf3d7779a3a8ec954d8ec0fb8a10f24982" ], "author": { "name": "Matthieu Baerts (NGI0)", "email": "matttbe@kernel.org", "time": "Sat Jul 27 12:01:25 2024 +0200" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Tue Jul 30 10:27:29 2024 +0200" }, "message": "mptcp: pm: only set request_bkup flag when sending MP_PRIO\n\nThe \u0027backup\u0027 flag from mptcp_subflow_context structure is supposed to be\nset only when the other peer flagged a subflow as backup, not the\nopposite.\n\nFixes: 067065422fcd (\"mptcp: add the outgoing MP_PRIO support\")\nCc: stable@vger.kernel.org\nReviewed-by: Mat Martineau \u003cmartineau@kernel.org\u003e\nSigned-off-by: Matthieu Baerts (NGI0) \u003cmatttbe@kernel.org\u003e\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n\n" }, { "commit": "efd340bf3d7779a3a8ec954d8ec0fb8a10f24982", "tree": "54a61da08972b3a15e07e9cb7014bb7b49c49d5d", "parents": [ "b6a66e521a2032f7fcba2af5a9bcbaeaa19b7ca3" ], "author": { "name": "Matthieu Baerts (NGI0)", "email": "matttbe@kernel.org", "time": "Sat Jul 27 12:01:24 2024 +0200" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Tue Jul 30 10:27:29 2024 +0200" }, "message": "mptcp: distinguish rcv vs sent backup flag in requests\n\nWhen sending an MP_JOIN + SYN + ACK, it is possible to mark the subflow\nas \u0027backup\u0027 by setting the flag with the same name. Before this patch,\nthe backup was set if the other peer set it in its MP_JOIN + SYN\nrequest.\n\nIt is not correct: the backup flag should be set in the MPJ+SYN+ACK only\nif the host asks for it, and not mirroring what was done by the other\npeer. It is then required to have a dedicated bit for each direction,\nsimilar to what is done in the subflow context.\n\nFixes: f296234c98a8 (\"mptcp: Add handling of incoming MP_JOIN requests\")\nCc: stable@vger.kernel.org\nReviewed-by: Mat Martineau \u003cmartineau@kernel.org\u003e\nSigned-off-by: Matthieu Baerts (NGI0) \u003cmatttbe@kernel.org\u003e\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n\n" }, { "commit": "b6a66e521a2032f7fcba2af5a9bcbaeaa19b7ca3", "tree": "244ec621cec59423108dcae43814a5ac856e3407", "parents": [ "039564d2fd37b122ec0d268e2ee6334e7169e225" ], "author": { "name": "Matthieu Baerts (NGI0)", "email": "matttbe@kernel.org", "time": "Sat Jul 27 12:01:23 2024 +0200" }, "committer": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Tue Jul 30 10:27:29 2024 +0200" }, "message": "mptcp: sched: check both directions for backup\n\nThe \u0027mptcp_subflow_context\u0027 structure has two items related to the\nbackup flags:\n\n - \u0027backup\u0027: the subflow has been marked as backup by the other peer\n\n - \u0027request_bkup\u0027: the backup flag has been set by the host\n\nBefore this patch, the scheduler was only looking at the \u0027backup\u0027 flag.\nThat can make sense in some cases, but it looks like that\u0027s not what we\nwanted for the general use, because either the path-manager was setting\nboth of them when sending an MP_PRIO, or the receiver was duplicating\nthe \u0027backup\u0027 flag in the subflow request.\n\nNote that the use of these two flags in the path-manager are going to be\nfixed in the next commits, but this change here is needed not to modify\nthe behaviour.\n\nFixes: f296234c98a8 (\"mptcp: Add handling of incoming MP_JOIN requests\")\nCc: stable@vger.kernel.org\nReviewed-by: Mat Martineau \u003cmartineau@kernel.org\u003e\nSigned-off-by: Matthieu Baerts (NGI0) \u003cmatttbe@kernel.org\u003e\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\n\n" }, { "commit": "94ede2a3e9135764736221c080ac7c0ad993dc2d", "tree": "10ce1d5ebc4c113a1c951d67395e14f517a2c578", "parents": [ "6b5faec9f564ea627c66064a4a6a5904fe5a07dd" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jul 29 16:34:17 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jul 29 16:34:17 2024 -0700" }, "message": "profiling: remove stale percpu flip buffer variables\n\nFor some reason I didn\u0027t see this issue on my arm64 or x86-64 builds,\nbut Stephen Rothwell reports that commit 2accfdb7eff6 (\"profiling:\nattempt to remove per-cpu profile flip buffer\") left these static\nvariables around, and the powerpc build is unhappy about them:\n\n kernel/profile.c:52:28: warning: \u0027cpu_profile_flip\u0027 defined but not used [-Wunused-variable]\n 52 | static DEFINE_PER_CPU(int, cpu_profile_flip);\n | ^~~~~~~~~~~~~~~~\n ..\n\nSo remove these stale left-over remnants too.\n\nFixes: 2accfdb7eff6 (\"profiling: attempt to remove per-cpu profile flip buffer\")\nReported-by: Stephen Rothwell \u003csfr@canb.auug.org.au\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "41c24102af7b6236277a214428b203d51a3462df", "tree": "88899ab944af9bac21c481a0130ddc8277801cc2", "parents": [ "039564d2fd37b122ec0d268e2ee6334e7169e225" ], "author": { "name": "Stanislav Fomichev", "email": "sdf@fomichev.me", "time": "Thu Jul 25 14:40:29 2024 -0700" }, "committer": { "name": "Andrii Nakryiko", "email": "andrii@kernel.org", "time": "Mon Jul 29 13:43:31 2024 -0700" }, "message": "selftests/bpf: Filter out _GNU_SOURCE when compiling test_cpp\n\nJakub reports build failures when merging linux/master with net tree:\n\nCXX test_cpp\nIn file included from \u003cbuilt-in\u003e:454:\n\u003ccommand line\u003e:2:9: error: \u0027_GNU_SOURCE\u0027 macro redefined [-Werror,-Wmacro-redefined]\n 2 | #define _GNU_SOURCE\n | ^\n\u003cbuilt-in\u003e:445:9: note: previous definition is here\n 445 | #define _GNU_SOURCE 1\n\nThe culprit is commit cc937dad85ae (\"selftests: centralize -D_GNU_SOURCE\u003d to\nCFLAGS in lib.mk\") which unconditionally added -D_GNU_SOUCE to CLFAGS.\nApparently clang++ also unconditionally adds it for the C++ targets [0]\nwhich causes a conflict. Add small change in the selftests makefile\nto filter it out for test_cpp.\n\nNot sure which tree it should go via, targeting bpf for now, but net\nmight be better?\n\n0: https://stackoverflow.com/questions/11670581/why-is-gnu-source-defined-by-default-and-how-to-turn-it-off\n\nSigned-off-by: Stanislav Fomichev \u003csdf@fomichev.me\u003e\nSigned-off-by: Andrii Nakryiko \u003candrii@kernel.org\u003e\nAcked-by: Jiri Olsa \u003cjolsa@kernel.org\u003e\nLink: https://lore.kernel.org/bpf/20240725214029.1760809-1-sdf@fomichev.me\n" }, { "commit": "6b5faec9f564ea627c66064a4a6a5904fe5a07dd", "tree": "a6981158a16b59d7effbac1fdb2c887a293a364a", "parents": [ "10826505f52357c7c9e12358db97a3acbe82e958", "facdbdfe0e6202d74758387ae9189c39f7b4b16c" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jul 29 13:07:05 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jul 29 13:07:05 2024 -0700" }, "message": "Merge tag \u0027for-linus-2024072901\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/hid/hid\n\nPull HID fixes from Benjamin Tissoires:\n\n - fixes for HID-BPF after the merge with the bpf tree (Arnd Bergmann\n and Benjamin Tissoires)\n\n - some tool type fix for the Wacom driver (Tatsunosuke Tobita)\n\n - a reorder of the sensor discovery to ensure the HID AMD SFH is\n removed when no sensors are available (Basavaraj Natikar)\n\n* tag \u0027for-linus-2024072901\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/hid/hid:\n selftests/hid: add test for attaching multiple time the same struct_ops\n HID: bpf: prevent the same struct_ops to be attached more than once\n selftests/hid: disable struct_ops auto-attach\n selftests/hid: fix bpf_wq new API\n HID: amd_sfh: Move sensor discovery before HID device initialization\n hid: bpf: add BPF_JIT dependency\n HID: wacom: more appropriate tool type categorization\n HID: wacom: Modify pen IDs\n" }, { "commit": "10826505f52357c7c9e12358db97a3acbe82e958", "tree": "98d225aef51ab86a6ba784f05bc04af5e11966a6", "parents": [ "cec6937dd1aae1b38d147bd190cb895d06cf96d0", "6d834691da474ed1c648753d3d3a3ef8379fa1c1" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jul 29 12:53:37 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jul 29 12:53:37 2024 -0700" }, "message": "Merge tag \u0027for_linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost\n\nPull virtio fixes from Michael Tsirkin:\n \"The biggest thing here is the adminq change - but it looks like the\n only way to avoid headq blocking causing indefinite stalls.\n\n This fixes three issues:\n\n - Prevent admin commands on one VF blocking another.\n\n This prevents a bad VF from blocking a good one, as well as fixing\n a scalability issue with large # of VFs\n\n - Correctly return error on command failure on octeon. We used to\n treat failed commands as a success.\n\n - Fix modpost warning when building virtio_dma_buf. Harmless, but the\n fix is trivial\"\n\n* tag \u0027for_linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost:\n virtio_pci_modern: remove admin queue serialization lock\n virtio_pci_modern: use completion instead of busy loop to wait on admin cmd result\n virtio_pci_modern: pass cmd as an identification token\n virtio_pci_modern: create admin queue of queried size\n virtio: create admin queues alongside other virtqueues\n virtio_pci: pass vq info as an argument to vp_setup_vq()\n virtio: push out code to vp_avq_index()\n virtio_pci_modern: treat vp_dev-\u003eadmin_vq.info.vq pointer as static\n virtio_pci: introduce vector allocation fallback for slow path virtqueues\n virtio_pci: pass vector policy enum to vp_find_one_vq_msix()\n virtio_pci: pass vector policy enum to vp_find_vqs_msix()\n virtio_pci: simplify vp_request_msix_vectors() call a bit\n virtio_pci: push out single vq find code to vp_find_one_vq_msix()\n vdpa/octeon_ep: Fix error code in octep_process_mbox()\n virtio: add missing MODULE_DESCRIPTION() macro\n" }, { "commit": "cec6937dd1aae1b38d147bd190cb895d06cf96d0", "tree": "971a172521f4d0d839a3737a8046b8a43c7d7601", "parents": [ "2accfdb7eff65f390c4308b0e9cb7c3fe48ad63c" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jul 29 12:05:06 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jul 29 12:05:06 2024 -0700" }, "message": "task_work: make TWA_NMI_CURRENT handling conditional on IRQ_WORK\n\nThe TWA_NMI_CURRENT handling very much depends on IRQ_WORK, but that\nisn\u0027t universally enabled everywhere.\n\nMaybe the IRQ_WORK infrastructure should just be unconditional - x86\nends up indirectly enabling it through unconditionally enabling\nPERF_EVENTS, for example. But it also gets enabled by having SMP\nsupport, or even if you just have PRINTK enabled.\n\nBut in the meantime TWA_NMI_CURRENT causes tons of build failures on\nvarious odd minimal configs. Which did show up in linux-next, but\ndespite that nobody bothered to fix it or even inform me until -rc1 was\nout.\n\nFixes: 466e4d801cd4 (\"task_work: Add TWA_NMI_CURRENT as an additional notify mode\")\nReported-by: Naresh Kamboju \u003cnaresh.kamboju@linaro.org\u003e\nReported-by: kernelci.org bot \u003cbot@kernelci.org\u003e\nReported-by: Guenter Roeck \u003clinux@roeck-us.net\u003e\nCc: Peter Zijlstra \u003cpeterz@infradead.org\u003e\nCc: Sebastian Andrzej Siewior \u003cbigeasy@linutronix.de\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "2accfdb7eff65f390c4308b0e9cb7c3fe48ad63c", "tree": "17eadc8c40ec82aab8c46932191d7af39d2469d1", "parents": [ "7c51f7bbf057f82aeba3390c39ef61b244181c09" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jul 29 10:58:28 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jul 29 10:58:28 2024 -0700" }, "message": "profiling: attempt to remove per-cpu profile flip buffer\n\nThis is the really old legacy kernel profiling code, which has long\nsince been obviated by \"real profiling\" (ie \u0027prof\u0027 and company), and\nmainly remains as a source of syzbot reports.\n\nThere are anecdotal reports that people still use it for boot-time\nprofiling, but it\u0027s unlikely that such use would care about the old NUMA\noptimizations in this code from 2004 (commit ad02973d42: \"profile: 512x\nAltix timer interrupt livelock fix\" in the BK import archive at [1])\n\nSo in order to head off future syzbot reports, let\u0027s try to simplify\nthis code and get rid of the per-cpu profile buffers that are quite a\nlarge portion of the complexity footprint of this thing (including CPU\nhotplug callbacks etc).\n\nIt\u0027s unlikely anybody will actually notice, or possibly, as Thomas put\nit: \"Only people who indulge in nostalgia will notice :)\".\n\nThat said, if it turns out that this code is actually actively used by\nsomebody, we can always revert this removal. Thus the \"attempt\" in the\nsummary line.\n\n[ Note: in a small nod to \"the profiling code can cause NUMA problems\",\n this also removes the \"increment the last entry in the profiling array\n on any unknown hits\" logic. That would account any program counter in\n a module to that single counter location, and might exacerbate any\n NUMA cacheline bouncing issues ]\n\nLink: https://lore.kernel.org/all/CAHk-\u003dwgs52BxT4Zjmjz8aNvHWKxf5_ThBY4bYL1Y6CTaNL2dTw@mail.gmail.com/\nLink: https://git.kernel.org/pub/scm/linux/kernel/git/tglx/history.git [1]\nCc: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nCc: Tetsuo Handa \u003cpenguin-kernel@i-love.sakura.ne.jp\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "7c51f7bbf057f82aeba3390c39ef61b244181c09", "tree": "07b2357f8dc51dbab1d8de4ad5f55a0ce2910ea3", "parents": [ "99d3bf5f7377d42f8be60a6b9cb60fb0be34dceb" ], "author": { "name": "Tetsuo Handa", "email": "penguin-kernel@I-love.SAKURA.ne.jp", "time": "Sat Jul 27 19:59:57 2024 +0900" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jul 29 10:45:54 2024 -0700" }, "message": "profiling: remove prof_cpu_mask\n\nsyzbot is reporting uninit-value at profile_hits(), for there is a race\nwindow between\n\n if (!alloc_cpumask_var(\u0026prof_cpu_mask, GFP_KERNEL))\n return -ENOMEM;\n cpumask_copy(prof_cpu_mask, cpu_possible_mask);\n\nin profile_init() and\n\n cpumask_available(prof_cpu_mask) \u0026\u0026\n cpumask_test_cpu(smp_processor_id(), prof_cpu_mask))\n\nin profile_tick(); prof_cpu_mask remains uninitialzed until cpumask_copy()\ncompletes while cpumask_available(prof_cpu_mask) returns true as soon as\nalloc_cpumask_var(\u0026prof_cpu_mask) completes.\n\nWe could replace alloc_cpumask_var() with zalloc_cpumask_var() and\ncall cpumask_copy() from create_proc_profile() on only UP kernels, for\nprofile_online_cpu() calls cpumask_set_cpu() as needed via\ncpuhp_setup_state(CPUHP_AP_ONLINE_DYN) on SMP kernels. But this patch\nremoves prof_cpu_mask because it seems unnecessary.\n\nThe cpumask_test_cpu(smp_processor_id(), prof_cpu_mask) test\nin profile_tick() is likely always true due to\n\n a CPU cannot call profile_tick() if that CPU is offline\n\nand\n\n cpumask_set_cpu(cpu, prof_cpu_mask) is called when that CPU becomes\n online and cpumask_clear_cpu(cpu, prof_cpu_mask) is called when that\n CPU becomes offline\n\n. This test could be false during transition between online and offline.\n\nBut according to include/linux/cpuhotplug.h , CPUHP_PROFILE_PREPARE\nbelongs to PREPARE section, which means that the CPU subjected to\nprofile_dead_cpu() cannot be inside profile_tick() (i.e. no risk of\nuse-after-free bug) because interrupt for that CPU is disabled during\nPREPARE section. Therefore, this test is guaranteed to be true, and\ncan be removed. (Since profile_hits() checks prof_buffer !\u003d NULL, we\ndon\u0027t need to check prof_buffer !\u003d NULL here unless get_irq_regs() or\nuser_mode() is such slow that we want to avoid when prof_buffer \u003d\u003d NULL).\n\ndo_profile_hits() is called from profile_tick() from timer interrupt\nonly if cpumask_test_cpu(smp_processor_id(), prof_cpu_mask) is true and\nprof_buffer is not NULL. But syzbot is also reporting that sometimes\ndo_profile_hits() is called while current thread is still doing vzalloc(),\nwhere prof_buffer must be NULL at this moment. This indicates that multiple\nthreads concurrently tried to write to /sys/kernel/profiling interface,\nwhich caused that somebody else try to re-allocate prof_buffer despite\nsomebody has already allocated prof_buffer. Fix this by using\nserialization.\n\nReported-by: syzbot \u003csyzbot+b1a83ab2a9eb9321fbdd@syzkaller.appspotmail.com\u003e\nCloses: https://syzkaller.appspot.com/bug?extid\u003db1a83ab2a9eb9321fbdd\nSigned-off-by: Tetsuo Handa \u003cpenguin-kernel@I-love.SAKURA.ne.jp\u003e\nTested-by: syzbot \u003csyzbot+b1a83ab2a9eb9321fbdd@syzkaller.appspotmail.com\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "99d3bf5f7377d42f8be60a6b9cb60fb0be34dceb", "tree": "78dd9dff2065f2eaf5a9e981f84d56eed2346d10", "parents": [ "3894840a7a11aa06cc3b0d5a2d1b5f6878127903" ], "author": { "name": "Tetsuo Handa", "email": "penguin-kernel@I-love.SAKURA.ne.jp", "time": "Mon Jul 29 21:51:30 2024 +0900" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jul 29 10:44:48 2024 -0700" }, "message": "Input: MT - limit max slots\n\nsyzbot is reporting too large allocation at input_mt_init_slots(), for\nnum_slots is supplied from userspace using ioctl(UI_DEV_CREATE).\n\nSince nobody knows possible max slots, this patch chose 1024.\n\nReported-by: syzbot \u003csyzbot+0122fa359a69694395d5@syzkaller.appspotmail.com\u003e\nCloses: https://syzkaller.appspot.com/bug?extid\u003d0122fa359a69694395d5\nSuggested-by: Dmitry Torokhov \u003cdmitry.torokhov@gmail.com\u003e\nSigned-off-by: Tetsuo Handa \u003cpenguin-kernel@I-love.SAKURA.ne.jp\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "3894840a7a11aa06cc3b0d5a2d1b5f6878127903", "tree": "ff80cfca0133b54914f7c1bd571383d3276a2857", "parents": [ "dc1c8034e31b14a2e5e212104ec508aec44ce1b9", "d7ac67b94390945872713f9bb0c141bf158c83aa" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jul 29 10:33:51 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jul 29 10:33:51 2024 -0700" }, "message": "Merge tag \u0027for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/rmk/linux\n\nPull ARM updates from Russell King:\n\n - ftrace: don\u0027t assume stack frames are contiguous in memory\n\n - remove unused mod_inwind_map structure\n\n - spelling fixes\n\n - allow use of LD dead code/data elimination\n\n - fix callchain_trace() return value\n\n - add support for stackleak gcc plugin\n\n - correct some reset asm function prototypes for CFI\n\n[ Missed the merge window because Russell forgot to push out ]\n\n* tag \u0027for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/rmk/linux:\n ARM: 9408/1: mm: CFI: Fix some erroneous reset prototypes\n ARM: 9407/1: Add support for STACKLEAK gcc plugin\n ARM: 9406/1: Fix callchain_trace() return value\n ARM: 9404/1: arm32: enable HAVE_LD_DEAD_CODE_DATA_ELIMINATION\n ARM: 9403/1: Alpine: Spelling s/initialiing/initializing/\n ARM: 9402/1: Kconfig: Spelling s/Cortex A-/Cortex-A/\n ARM: 9400/1: Remove unused struct \u0027mod_unwind_map\u0027\n" }, { "commit": "939b656bc8ab203fdbde26ccac22bcb7f0985be5", "tree": "8a02666571558de471818c21eae8eb46d39cc415", "parents": [ "8cd44dd1d17a23d5cc8c443c659ca57aa76e2fa5" ], "author": { "name": "Filipe Manana", "email": "fdmanana@suse.com", "time": "Fri Jul 26 11:12:52 2024 +0100" }, "committer": { "name": "David Sterba", "email": "dsterba@suse.com", "time": "Mon Jul 29 19:21:22 2024 +0200" }, "message": "btrfs: fix corruption after buffer fault in during direct IO append write\n\nDuring an append (O_APPEND write flag) direct IO write if the input buffer\nwas not previously faulted in, we can corrupt the file in a way that the\nfinal size is unexpected and it includes an unexpected hole.\n\nThe problem happens like this:\n\n1) We have an empty file, with size 0, for example;\n\n2) We do an O_APPEND direct IO with a length of 4096 bytes and the input\n buffer is not currently faulted in;\n\n3) We enter btrfs_direct_write(), lock the inode and call\n generic_write_checks(), which calls generic_write_checks_count(), and\n that function sets the iocb position to 0 with the following code:\n\n\tif (iocb-\u003eki_flags \u0026 IOCB_APPEND)\n\t\tiocb-\u003eki_pos \u003d i_size_read(inode);\n\n4) We call btrfs_dio_write() and enter into iomap, which will end up\n calling btrfs_dio_iomap_begin() and that calls\n btrfs_get_blocks_direct_write(), where we update the i_size of the\n inode to 4096 bytes;\n\n5) After btrfs_dio_iomap_begin() returns, iomap will attempt to access\n the page of the write input buffer (at iomap_dio_bio_iter(), with a\n call to bio_iov_iter_get_pages()) and fail with -EFAULT, which gets\n returned to btrfs at btrfs_direct_write() via btrfs_dio_write();\n\n6) At btrfs_direct_write() we get the -EFAULT error, unlock the inode,\n fault in the write buffer and then goto to the label \u0027relock\u0027;\n\n7) We lock again the inode, do all the necessary checks again and call\n again generic_write_checks(), which calls generic_write_checks_count()\n again, and there we set the iocb\u0027s position to 4K, which is the current\n i_size of the inode, with the following code pointed above:\n\n if (iocb-\u003eki_flags \u0026 IOCB_APPEND)\n iocb-\u003eki_pos \u003d i_size_read(inode);\n\n8) Then we go again to btrfs_dio_write() and enter iomap and the write\n succeeds, but it wrote to the file range [4K, 8K), leaving a hole in\n the [0, 4K) range and an i_size of 8K, which goes against the\n expectations of having the data written to the range [0, 4K) and get an\n i_size of 4K.\n\nFix this by not unlocking the inode before faulting in the input buffer,\nin case we get -EFAULT or an incomplete write, and not jumping to the\n\u0027relock\u0027 label after faulting in the buffer - instead jump to a location\nimmediately before calling iomap, skipping all the write checks and\nrelocking. This solves this problem and it\u0027s fine even in case the input\nbuffer is memory mapped to the same file range, since only holding the\nrange locked in the inode\u0027s io tree can cause a deadlock, it\u0027s safe to\nkeep the inode lock (VFS lock), as was fixed and described in commit\n51bd9563b678 (\"btrfs: fix deadlock due to page faults during direct IO\nreads and writes\").\n\nA sample reproducer provided by a reporter is the following:\n\n $ cat test.c\n #ifndef _GNU_SOURCE\n #define _GNU_SOURCE\n #endif\n\n #include \u003cfcntl.h\u003e\n #include \u003cstdio.h\u003e\n #include \u003csys/mman.h\u003e\n #include \u003csys/stat.h\u003e\n #include \u003cunistd.h\u003e\n\n int main(int argc, char *argv[])\n {\n if (argc \u003c 2) {\n fprintf(stderr, \"Usage: %s \u003ctest file\u003e\\n\", argv[0]);\n return 1;\n }\n\n int fd \u003d open(argv[1], O_WRONLY | O_CREAT | O_TRUNC | O_DIRECT |\n O_APPEND, 0644);\n if (fd \u003c 0) {\n perror(\"creating test file\");\n return 1;\n }\n\n char *buf \u003d mmap(NULL, 4096, PROT_READ,\n MAP_PRIVATE | MAP_ANONYMOUS, -1, 0);\n ssize_t ret \u003d write(fd, buf, 4096);\n if (ret \u003c 0) {\n perror(\"pwritev2\");\n return 1;\n }\n\n struct stat stbuf;\n ret \u003d fstat(fd, \u0026stbuf);\n if (ret \u003c 0) {\n perror(\"stat\");\n return 1;\n }\n\n printf(\"size: %llu\\n\", (unsigned long long)stbuf.st_size);\n return stbuf.st_size \u003d\u003d 4096 ? 0 : 1;\n }\n\nA test case for fstests will be sent soon.\n\nReported-by: Hanna Czenczek \u003chreitz@redhat.com\u003e\nLink: https://lore.kernel.org/linux-btrfs/0b841d46-12fe-4e64-9abb-871d8d0de271@redhat.com/\nFixes: 8184620ae212 (\"btrfs: fix lost file sync on direct IO write with nowait and dsync iocb\")\nCC: stable@vger.kernel.org # 6.1+\nTested-by: Hanna Czenczek \u003chreitz@redhat.com\u003e\nReviewed-by: Josef Bacik \u003cjosef@toxicpanda.com\u003e\nSigned-off-by: Filipe Manana \u003cfdmanana@suse.com\u003e\nSigned-off-by: David Sterba \u003cdsterba@suse.com\u003e\n" }, { "commit": "8cd44dd1d17a23d5cc8c443c659ca57aa76e2fa5", "tree": "0f3e91d595360d715d0d087a4413a1db817bdd61", "parents": [ "d89c285d28491d8f10534c262ac9e6bdcbe1b4d2" ], "author": { "name": "Naohiro Aota", "email": "naohiro.aota@wdc.com", "time": "Wed Feb 15 09:18:02 2023 +0900" }, "committer": { "name": "David Sterba", "email": "dsterba@suse.com", "time": "Mon Jul 29 19:21:19 2024 +0200" }, "message": "btrfs: zoned: fix zone_unusable accounting on making block group read-write again\n\nWhen btrfs makes a block group read-only, it adds all free regions in the\nblock group to space_info-\u003ebytes_readonly. That free space excludes\nreserved and pinned regions. OTOH, when btrfs makes the block group\nread-write again, it moves all the unused regions into the block group\u0027s\nzone_unusable. That unused region includes reserved and pinned regions.\nAs a result, it counts too much zone_unusable bytes.\n\nFortunately (or unfortunately), having erroneous zone_unusable does not\naffect the calculation of space_info-\u003ebytes_readonly, because free\nspace (num_bytes in btrfs_dec_block_group_ro) calculation is done based on\nthe erroneous zone_unusable and it reduces the num_bytes just to cancel the\nerror.\n\nThis behavior can be easily discovered by adding a WARN_ON to check e.g,\n\"bg-\u003epinned \u003e 0\" in btrfs_dec_block_group_ro(), and running fstests test\ncase like btrfs/282.\n\nFix it by properly considering pinned and reserved in\nbtrfs_dec_block_group_ro(). Also, add a WARN_ON and introduce\nbtrfs_space_info_update_bytes_zone_unusable() to catch a similar mistake.\n\nFixes: 169e0da91a21 (\"btrfs: zoned: track unusable bytes for zones\")\nCC: stable@vger.kernel.org # 5.15+\nSigned-off-by: Naohiro Aota \u003cnaohiro.aota@wdc.com\u003e\nReviewed-by: Josef Bacik \u003cjosef@toxicpanda.com\u003e\nReviewed-by: Johannes Thumshirn \u003cjohannes.thumshirn@wdc.com\u003e\nSigned-off-by: David Sterba \u003cdsterba@suse.com\u003e\n" }, { "commit": "d89c285d28491d8f10534c262ac9e6bdcbe1b4d2", "tree": "65e74d64790fb797fa3507f86e39f7aa691786f6", "parents": [ "478574370bef7951fbd9ef5155537d6cbed49472" ], "author": { "name": "Naohiro Aota", "email": "naohiro.aota@wdc.com", "time": "Thu Jul 11 23:50:58 2024 +0900" }, "committer": { "name": "David Sterba", "email": "dsterba@suse.com", "time": "Mon Jul 29 19:21:04 2024 +0200" }, "message": "btrfs: do not subtract delalloc from avail bytes\n\nThe block group\u0027s avail bytes printed when dumping a space info subtract\nthe delalloc_bytes. However, as shown in btrfs_add_reserved_bytes() and\nbtrfs_free_reserved_bytes(), it is added or subtracted along with\n\"reserved\" for the delalloc case, which means the \"delalloc_bytes\" is a\npart of the \"reserved\" bytes. So, excluding it to calculate the avail space\ncounts delalloc_bytes twice, which can lead to an invalid result.\n\nFixes: e50b122b832b (\"btrfs: print available space for a block group when dumping a space info\")\nCC: stable@vger.kernel.org # 6.6+\nSigned-off-by: Naohiro Aota \u003cnaohiro.aota@wdc.com\u003e\nReviewed-by: Boris Burkov \u003cboris@bur.io\u003e\nSigned-off-by: David Sterba \u003cdsterba@suse.com\u003e\n" }, { "commit": "478574370bef7951fbd9ef5155537d6cbed49472", "tree": "c5c69d24a4b3a4db0e0910b88bf130c54bb1f806", "parents": [ "de9f46cb0044a9b9f825d7695ae235863461dc00" ], "author": { "name": "Boris Burkov", "email": "boris@bur.io", "time": "Mon Jul 22 16:49:45 2024 -0700" }, "committer": { "name": "David Sterba", "email": "dsterba@suse.com", "time": "Mon Jul 29 19:20:51 2024 +0200" }, "message": "btrfs: make cow_file_range_inline() honor locked_page on error\n\nThe btrfs buffered write path runs through __extent_writepage() which\nhas some tricky return value handling for writepage_delalloc().\nSpecifically, when that returns 1, we exit, but for other return values\nwe continue and end up calling btrfs_folio_end_all_writers(). If the\nfolio has been unlocked (note that we check the PageLocked bit at the\nstart of __extent_writepage()), this results in an assert panic like\nthis one from syzbot:\n\n BTRFS: error (device loop0 state EAL) in free_log_tree:3267: errno\u003d-5 IO failure\n BTRFS warning (device loop0 state EAL): Skipping commit of aborted transaction.\n BTRFS: error (device loop0 state EAL) in cleanup_transaction:2018: errno\u003d-5 IO failure\n assertion failed: folio_test_locked(folio), in fs/btrfs/subpage.c:871\n ------------[ cut here ]------------\n kernel BUG at fs/btrfs/subpage.c:871!\n Oops: invalid opcode: 0000 [#1] PREEMPT SMP KASAN PTI\n CPU: 1 PID: 5090 Comm: syz-executor225 Not tainted\n 6.10.0-syzkaller-05505-gb1bc554e009e #0\n Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS\n Google 06/27/2024\n RIP: 0010:btrfs_folio_end_all_writers+0x55b/0x610 fs/btrfs/subpage.c:871\n Code: e9 d3 fb ff ff e8 25 22 c2 fd 48 c7 c7 c0 3c 0e 8c 48 c7 c6 80 3d\n 0e 8c 48 c7 c2 60 3c 0e 8c b9 67 03 00 00 e8 66 47 ad 07 90 \u003c0f\u003e 0b e8\n 6e 45 b0 07 4c 89 ff be 08 00 00 00 e8 21 12 25 fe 4c 89\n RSP: 0018:ffffc900033d72e0 EFLAGS: 00010246\n RAX: 0000000000000045 RBX: 00fff0000000402c RCX: 663b7a08c50a0a00\n RDX: 0000000000000000 RSI: 0000000080000000 RDI: 0000000000000000\n RBP: ffffc900033d73b0 R08: ffffffff8176b98c R09: 1ffff9200067adfc\n R10: dffffc0000000000 R11: fffff5200067adfd R12: 0000000000000001\n R13: dffffc0000000000 R14: 0000000000000000 R15: ffffea0001cbee80\n FS: 0000000000000000(0000) GS:ffff8880b9500000(0000)\n knlGS:0000000000000000\n CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033\n CR2: 00007f5f076012f8 CR3: 000000000e134000 CR4: 00000000003506f0\n DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000\n DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400\n Call Trace:\n \u003cTASK\u003e\n __extent_writepage fs/btrfs/extent_io.c:1597 [inline]\n extent_write_cache_pages fs/btrfs/extent_io.c:2251 [inline]\n btrfs_writepages+0x14d7/0x2760 fs/btrfs/extent_io.c:2373\n do_writepages+0x359/0x870 mm/page-writeback.c:2656\n filemap_fdatawrite_wbc+0x125/0x180 mm/filemap.c:397\n __filemap_fdatawrite_range mm/filemap.c:430 [inline]\n __filemap_fdatawrite mm/filemap.c:436 [inline]\n filemap_flush+0xdf/0x130 mm/filemap.c:463\n btrfs_release_file+0x117/0x130 fs/btrfs/file.c:1547\n __fput+0x24a/0x8a0 fs/file_table.c:422\n task_work_run+0x24f/0x310 kernel/task_work.c:222\n exit_task_work include/linux/task_work.h:40 [inline]\n do_exit+0xa2f/0x27f0 kernel/exit.c:877\n do_group_exit+0x207/0x2c0 kernel/exit.c:1026\n __do_sys_exit_group kernel/exit.c:1037 [inline]\n __se_sys_exit_group kernel/exit.c:1035 [inline]\n __x64_sys_exit_group+0x3f/0x40 kernel/exit.c:1035\n x64_sys_call+0x2634/0x2640\n arch/x86/include/generated/asm/syscalls_64.h:232\n do_syscall_x64 arch/x86/entry/common.c:52 [inline]\n do_syscall_64+0xf3/0x230 arch/x86/entry/common.c:83\n entry_SYSCALL_64_after_hwframe+0x77/0x7f\n RIP: 0033:0x7f5f075b70c9\n Code: Unable to access opcode bytes at\n 0x7f5f075b709f.\n\nI was hitting the same issue by doing hundreds of accelerated runs of\ngeneric/475, which also hits IO errors by design.\n\nI instrumented that reproducer with bpftrace and found that the\nundesirable folio_unlock was coming from the following callstack:\n\n folio_unlock+5\n __process_pages_contig+475\n cow_file_range_inline.constprop.0+230\n cow_file_range+803\n btrfs_run_delalloc_range+566\n writepage_delalloc+332\n __extent_writepage # inlined in my stacktrace, but I added it here\n extent_write_cache_pages+622\n\nLooking at the bisected-to patch in the syzbot report, Josef realized\nthat the logic of the cow_file_range_inline error path subtly changing.\nIn the past, on error, it jumped to out_unlock in cow_file_range(),\nwhich honors the locked_page, so when we ultimately call\nfolio_end_all_writers(), the folio of interest is still locked. After\nthe change, we always unlocked ignoring the locked_page, on both success\nand error. On the success path, this all results in returning 1 to\n__extent_writepage(), which skips the folio_end_all_writers() call,\nwhich makes it OK to have unlocked.\n\nFix the bug by wiring the locked_page into cow_file_range_inline() and\nonly setting locked_page to NULL on success.\n\nReported-by: syzbot+a14d8ac9af3a2a4fd0c8@syzkaller.appspotmail.com\nFixes: 0586d0a89e77 (\"btrfs: move extent bit and page cleanup into cow_file_range_inline\")\nCC: stable@vger.kernel.org # 6.10+\nReviewed-by: Qu Wenruo \u003cwqu@suse.com\u003e\nSigned-off-by: Boris Burkov \u003cboris@bur.io\u003e\nSigned-off-by: David Sterba \u003cdsterba@suse.com\u003e\n" }, { "commit": "963fb4612295a5c35b1b89c8bff3bdd4f9127af6", "tree": "c7a9e3489c5b1320680418d2bf6e4ba61e57c8e6", "parents": [ "6044ca26210ba72b3dcc649fae1cbedd9e6ab018" ], "author": { "name": "Maciej Fijalkowski", "email": "maciej.fijalkowski@intel.com", "time": "Fri Jul 26 20:17:16 2024 +0200" }, "committer": { "name": "Tony Nguyen", "email": "anthony.l.nguyen@intel.com", "time": "Mon Jul 29 08:52:29 2024 -0700" }, "message": "ice: xsk: fix txq interrupt mapping\n\nice_cfg_txq_interrupt() internally handles XDP Tx ring. Do not use\nice_for_each_tx_ring() in ice_qvec_cfg_msix() as this causing us to\ntreat XDP ring that belongs to queue vector as Tx ring and therefore\nmisconfiguring the interrupts.\n\nFixes: 2d4238f55697 (\"ice: Add support for AF_XDP\")\nReviewed-by: Shannon Nelson \u003cshannon.nelson@amd.com\u003e\nTested-by: Chandan Kumar Rout \u003cchandanx.rout@intel.com\u003e (A Contingent Worker at Intel)\nSigned-off-by: Maciej Fijalkowski \u003cmaciej.fijalkowski@intel.com\u003e\nSigned-off-by: Tony Nguyen \u003canthony.l.nguyen@intel.com\u003e\n" }, { "commit": "6044ca26210ba72b3dcc649fae1cbedd9e6ab018", "tree": "733df68fc63cce60237df04fc7e09943f75879bd", "parents": [ "ebc33a3f8d0aeddf19fd5827add24b82ae171829" ], "author": { "name": "Maciej Fijalkowski", "email": "maciej.fijalkowski@intel.com", "time": "Fri Jul 26 20:17:15 2024 +0200" }, "committer": { "name": "Tony Nguyen", "email": "anthony.l.nguyen@intel.com", "time": "Mon Jul 29 08:52:29 2024 -0700" }, "message": "ice: add missing WRITE_ONCE when clearing ice_rx_ring::xdp_prog\n\nIt is read by data path and modified from process context on remote cpu\nso it is needed to use WRITE_ONCE to clear the pointer.\n\nFixes: efc2214b6047 (\"ice: Add support for XDP\")\nReviewed-by: Shannon Nelson \u003cshannon.nelson@amd.com\u003e\nTested-by: Chandan Kumar Rout \u003cchandanx.rout@intel.com\u003e (A Contingent Worker at Intel)\nSigned-off-by: Maciej Fijalkowski \u003cmaciej.fijalkowski@intel.com\u003e\nSigned-off-by: Tony Nguyen \u003canthony.l.nguyen@intel.com\u003e\n" }, { "commit": "ebc33a3f8d0aeddf19fd5827add24b82ae171829", "tree": "e8c1a5d320752997cc73df853be09098c04ef672", "parents": [ "9da75a511c5558fa3da56759984fd1fa859186f0" ], "author": { "name": "Maciej Fijalkowski", "email": "maciej.fijalkowski@intel.com", "time": "Fri Jul 26 20:17:14 2024 +0200" }, "committer": { "name": "Tony Nguyen", "email": "anthony.l.nguyen@intel.com", "time": "Mon Jul 29 08:52:29 2024 -0700" }, "message": "ice: improve updating ice_{t,r}x_ring::xsk_pool\n\nxsk_buff_pool pointers that ice ring structs hold are updated via\nndo_bpf that is executed in process context while it can be read by\nremote CPU at the same time within NAPI poll. Use synchronize_net()\nafter pointer update and {READ,WRITE}_ONCE() when working with mentioned\npointer.\n\nFixes: 2d4238f55697 (\"ice: Add support for AF_XDP\")\nReviewed-by: Shannon Nelson \u003cshannon.nelson@amd.com\u003e\nTested-by: Chandan Kumar Rout \u003cchandanx.rout@intel.com\u003e (A Contingent Worker at Intel)\nSigned-off-by: Maciej Fijalkowski \u003cmaciej.fijalkowski@intel.com\u003e\nSigned-off-by: Tony Nguyen \u003canthony.l.nguyen@intel.com\u003e\n" }, { "commit": "9da75a511c5558fa3da56759984fd1fa859186f0", "tree": "29c3ce9391973504cbc7f888663ed6aad620f4e1", "parents": [ "d5922717994911e8f0eab736f3ba0d968c158823" ], "author": { "name": "Maciej Fijalkowski", "email": "maciej.fijalkowski@intel.com", "time": "Fri Jul 26 20:17:13 2024 +0200" }, "committer": { "name": "Tony Nguyen", "email": "anthony.l.nguyen@intel.com", "time": "Mon Jul 29 08:52:29 2024 -0700" }, "message": "ice: toggle netif_carrier when setting up XSK pool\n\nThis so we prevent Tx timeout issues. One of conditions checked on\nrunning in the background dev_watchdog() is netif_carrier_ok(), so let\nus turn it off when we disable the queues that belong to a q_vector\nwhere XSK pool is being configured. Turn carrier on in ice_qp_ena()\nonly when ice_get_link_status() tells us that physical link is up.\n\nFixes: 2d4238f55697 (\"ice: Add support for AF_XDP\")\nReviewed-by: Shannon Nelson \u003cshannon.nelson@amd.com\u003e\nTested-by: Chandan Kumar Rout \u003cchandanx.rout@intel.com\u003e (A Contingent Worker at Intel)\nSigned-off-by: Maciej Fijalkowski \u003cmaciej.fijalkowski@intel.com\u003e\nSigned-off-by: Tony Nguyen \u003canthony.l.nguyen@intel.com\u003e\n" }, { "commit": "d5922717994911e8f0eab736f3ba0d968c158823", "tree": "cbddecc20715a80cececc165701dbe35a1ce69ab", "parents": [ "405d9999aa0b4ae467ef391d1d9c7e0d30ad0841" ], "author": { "name": "Maciej Fijalkowski", "email": "maciej.fijalkowski@intel.com", "time": "Fri Jul 26 20:17:12 2024 +0200" }, "committer": { "name": "Tony Nguyen", "email": "anthony.l.nguyen@intel.com", "time": "Mon Jul 29 08:52:29 2024 -0700" }, "message": "ice: modify error handling when setting XSK pool in ndo_bpf\n\nDon\u0027t bail out right when spotting an error within ice_qp_{dis,ena}()\nbut rather track error and go through whole flow of disabling and\nenabling queue pair.\n\nFixes: 2d4238f55697 (\"ice: Add support for AF_XDP\")\nReviewed-by: Shannon Nelson \u003cshannon.nelson@amd.com\u003e\nTested-by: Chandan Kumar Rout \u003cchandanx.rout@intel.com\u003e (A Contingent Worker at Intel)\nSigned-off-by: Maciej Fijalkowski \u003cmaciej.fijalkowski@intel.com\u003e\nSigned-off-by: Tony Nguyen \u003canthony.l.nguyen@intel.com\u003e\n" }, { "commit": "405d9999aa0b4ae467ef391d1d9c7e0d30ad0841", "tree": "c2c0da42ce0eae01cb0976006d44131eaedc41a6", "parents": [ "1ff72a2f67791cd4ddad19ed830445f57b30e992" ], "author": { "name": "Maciej Fijalkowski", "email": "maciej.fijalkowski@intel.com", "time": "Fri Jul 26 20:17:11 2024 +0200" }, "committer": { "name": "Tony Nguyen", "email": "anthony.l.nguyen@intel.com", "time": "Mon Jul 29 08:52:29 2024 -0700" }, "message": "ice: replace synchronize_rcu with synchronize_net\n\nGiven that ice_qp_dis() is called under rtnl_lock, synchronize_net() can\nbe called instead of synchronize_rcu() so that XDP rings can finish its\njob in a faster way. Also let us do this as earlier in XSK queue disable\nflow.\n\nAdditionally, turn off regular Tx queue before disabling irqs and NAPI.\n\nFixes: 2d4238f55697 (\"ice: Add support for AF_XDP\")\nReviewed-by: Shannon Nelson \u003cshannon.nelson@amd.com\u003e\nTested-by: Chandan Kumar Rout \u003cchandanx.rout@intel.com\u003e (A Contingent Worker at Intel)\nSigned-off-by: Maciej Fijalkowski \u003cmaciej.fijalkowski@intel.com\u003e\nSigned-off-by: Tony Nguyen \u003canthony.l.nguyen@intel.com\u003e\n" }, { "commit": "1ff72a2f67791cd4ddad19ed830445f57b30e992", "tree": "71759c44d271e2b3b682c9c4a499eaedefcd95b1", "parents": [ "ec145a18687fec8dd97eeb4f30057fa4debef577" ], "author": { "name": "Maciej Fijalkowski", "email": "maciej.fijalkowski@intel.com", "time": "Fri Jul 26 20:17:10 2024 +0200" }, "committer": { "name": "Tony Nguyen", "email": "anthony.l.nguyen@intel.com", "time": "Mon Jul 29 08:52:29 2024 -0700" }, "message": "ice: don\u0027t busy wait for Rx queue disable in ice_qp_dis()\n\nWhen ice driver is spammed with multiple xdpsock instances and flow\ncontrol is enabled, there are cases when Rx queue gets stuck and unable\nto reflect the disable state in QRX_CTRL register. Similar issue has\npreviously been addressed in commit 13a6233b033f (\"ice: Add support to\nenable/disable all Rx queues before waiting\").\n\nTo workaround this, let us simply not wait for a disabled state as later\npatch will make sure that regardless of the encountered error in the\nprocess of disabling a queue pair, the Rx queue will be enabled.\n\nFixes: 2d4238f55697 (\"ice: Add support for AF_XDP\")\nReviewed-by: Shannon Nelson \u003cshannon.nelson@amd.com\u003e\nTested-by: Chandan Kumar Rout \u003cchandanx.rout@intel.com\u003e (A Contingent Worker at Intel)\nSigned-off-by: Maciej Fijalkowski \u003cmaciej.fijalkowski@intel.com\u003e\nSigned-off-by: Tony Nguyen \u003canthony.l.nguyen@intel.com\u003e\n" }, { "commit": "ec145a18687fec8dd97eeb4f30057fa4debef577", "tree": "8df88f7bbb44839a641a008fa754f7e3e3b68b58", "parents": [ "039564d2fd37b122ec0d268e2ee6334e7169e225" ], "author": { "name": "Michal Kubiak", "email": "michal.kubiak@intel.com", "time": "Fri Jul 26 20:17:09 2024 +0200" }, "committer": { "name": "Tony Nguyen", "email": "anthony.l.nguyen@intel.com", "time": "Mon Jul 29 08:52:29 2024 -0700" }, "message": "ice: respect netif readiness in AF_XDP ZC related ndo\u0027s\n\nAddress a scenario in which XSK ZC Tx produces descriptors to XDP Tx\nring when link is either not yet fully initialized or process of\nstopping the netdev has already started. To avoid this, add checks\nagainst carrier readiness in ice_xsk_wakeup() and in ice_xmit_zc().\nOne could argue that bailing out early in ice_xsk_wakeup() would be\nsufficient but given the fact that we produce Tx descriptors on behalf\nof NAPI that is triggered for Rx traffic, the latter is also needed.\n\nBringing link up is an asynchronous event executed within\nice_service_task so even though interface has been brought up there is\nstill a time frame where link is not yet ok.\n\nWithout this patch, when AF_XDP ZC Tx is used simultaneously with stack\nTx, Tx timeouts occur after going through link flap (admin brings\ninterface down then up again). HW seem to be unable to transmit\ndescriptor to the wire after HW tail register bump which in turn causes\nbit __QUEUE_STATE_STACK_XOFF to be set forever as\nnetdev_tx_completed_queue() sees no cleaned bytes on the input.\n\nFixes: 126cdfe1007a (\"ice: xsk: Improve AF_XDP ZC Tx and use batching API\")\nFixes: 2d4238f55697 (\"ice: Add support for AF_XDP\")\nReviewed-by: Shannon Nelson \u003cshannon.nelson@amd.com\u003e\nTested-by: Chandan Kumar Rout \u003cchandanx.rout@intel.com\u003e (A Contingent Worker at Intel)\nSigned-off-by: Michal Kubiak \u003cmichal.kubiak@intel.com\u003e\nSigned-off-by: Maciej Fijalkowski \u003cmaciej.fijalkowski@intel.com\u003e\nSigned-off-by: Tony Nguyen \u003canthony.l.nguyen@intel.com\u003e\n" }, { "commit": "039564d2fd37b122ec0d268e2ee6334e7169e225", "tree": "402867947e6b51f15027dafd583d0fc5fc09fcd1", "parents": [ "9415d375d8520e0ed55f0c0b058928da9a5b5b3d", "7c70bcc2a84cf925f655ea1ac4b8088062b144a3" ], "author": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Mon Jul 29 13:31:28 2024 +0100" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Mon Jul 29 13:31:28 2024 +0100" }, "message": "Merge branch \u0027mptcp-endpoint-readd-fixes\u0027 into main\n\nMatthieu Baerts says:\n\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\nmptcp: fix signal endpoint readd\n\nIssue #501 [1] showed that the Netlink PM currently doesn\u0027t correctly\nsupport removal and re-add of signal endpoints.\n\nPatches 1 and 2 address the issue: the first one in the userspace path-\nmanager, introduced in v5.19 ; and the second one in the in-kernel path-\nmanager, introduced in v5.7.\n\nPatch 3 introduces a related selftest. There is no \u0027Fixes\u0027 tag, because\nit might be hard to backport it automatically, as missing helpers in\nBash will not be caught when compiling the kernel or the selftests.\n\nThe last two patches address two small issues in the MPTCP selftests,\none introduced in v6.6., and the other one in v5.17.\n\nLink: https://github.com/multipath-tcp/mptcp_net-next/issues/501 [1]\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\nSigned-off-by: Matthieu Baerts (NGI0) \u003cmatttbe@kernel.org\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "7c70bcc2a84cf925f655ea1ac4b8088062b144a3", "tree": "402867947e6b51f15027dafd583d0fc5fc09fcd1", "parents": [ "4a2f48992ddf4b8c2fba846c6754089edae6db5a" ], "author": { "name": "Liu Jing", "email": "liujing@cmss.chinamobile.com", "time": "Sat Jul 27 11:04:03 2024 +0200" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Mon Jul 29 13:31:27 2024 +0100" }, "message": "selftests: mptcp: always close input\u0027s FD if opened\n\nIn main_loop_s function, when the open(cfg_input, O_RDONLY) function is\nrun, the last fd is not closed if the \"--cfg_repeat \u003e 0\" branch is not\ntaken.\n\nFixes: 05be5e273c84 (\"selftests: mptcp: add disconnect tests\")\nCc: stable@vger.kernel.org\nSigned-off-by: Liu Jing \u003cliujing@cmss.chinamobile.com\u003e\nReviewed-by: Matthieu Baerts (NGI0) \u003cmatttbe@kernel.org\u003e\nSigned-off-by: Matthieu Baerts (NGI0) \u003cmatttbe@kernel.org\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "4a2f48992ddf4b8c2fba846c6754089edae6db5a", "tree": "bf3522c02a69e57db42ad703cfd7d00469c1b0fa", "parents": [ "b5e2fb832f48bc01d937a053e0550a1465a2f05d" ], "author": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Sat Jul 27 11:04:02 2024 +0200" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Mon Jul 29 13:31:27 2024 +0100" }, "message": "selftests: mptcp: fix error path\n\npm_nl_check_endpoint() currently calls an not existing helper\nto mark the test as failed. Fix the wrong call.\n\nFixes: 03668c65d153 (\"selftests: mptcp: join: rework detailed report\")\nCc: stable@vger.kernel.org\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\nReviewed-by: Matthieu Baerts (NGI0) \u003cmatttbe@kernel.org\u003e\nSigned-off-by: Matthieu Baerts (NGI0) \u003cmatttbe@kernel.org\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "b5e2fb832f48bc01d937a053e0550a1465a2f05d", "tree": "0ec31760878b0f271d109d8aeb0a931c0cbfb83b", "parents": [ "4b317e0eb287bd30a1b329513531157c25e8b692" ], "author": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Sat Jul 27 11:04:01 2024 +0200" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Mon Jul 29 13:31:27 2024 +0100" }, "message": "selftests: mptcp: add explicit test case for remove/readd\n\nDelete and re-create a signal endpoint and ensure that the PM\nactually deletes and re-create the subflow.\n\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\nReviewed-by: Matthieu Baerts (NGI0) \u003cmatttbe@kernel.org\u003e\nSigned-off-by: Matthieu Baerts (NGI0) \u003cmatttbe@kernel.org\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "4b317e0eb287bd30a1b329513531157c25e8b692", "tree": "a55b5359a7bff0fae8c6b1cf6e8ab71059f3bd86", "parents": [ "167b93258d1e2230ee3e8a97669b4db4cc9e90aa" ], "author": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Sat Jul 27 11:04:00 2024 +0200" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Mon Jul 29 13:31:27 2024 +0100" }, "message": "mptcp: fix NL PM announced address accounting\n\nCurrently the per connection announced address counter is never\ndecreased. As a consequence, after connection establishment, if\nthe NL PM deletes an endpoint and adds a new/different one, no\nadditional subflow is created for the new endpoint even if the\ncurrent limits allow that.\n\nAddress the issue properly updating the signaled address counter\nevery time the NL PM removes such addresses.\n\nFixes: 01cacb00b35c (\"mptcp: add netlink-based PM\")\nCc: stable@vger.kernel.org\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\nReviewed-by: Matthieu Baerts (NGI0) \u003cmatttbe@kernel.org\u003e\nSigned-off-by: Matthieu Baerts (NGI0) \u003cmatttbe@kernel.org\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "167b93258d1e2230ee3e8a97669b4db4cc9e90aa", "tree": "7110f08aedc22be95a9a24b2de1b2db94e79141d", "parents": [ "9415d375d8520e0ed55f0c0b058928da9a5b5b3d" ], "author": { "name": "Paolo Abeni", "email": "pabeni@redhat.com", "time": "Sat Jul 27 11:03:59 2024 +0200" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Mon Jul 29 13:31:26 2024 +0100" }, "message": "mptcp: fix user-space PM announced address accounting\n\nCurrently the per-connection announced address counter is never\ndecreased. When the user-space PM is in use, this just affect\nthe information exposed via diag/sockopt, but it could still foul\nthe PM to wrong decision.\n\nAdd the missing accounting for the user-space PM\u0027s sake.\n\nFixes: 8b1c94da1e48 (\"mptcp: only send RM_ADDR in nl_cmd_remove\")\nCc: stable@vger.kernel.org\nSigned-off-by: Paolo Abeni \u003cpabeni@redhat.com\u003e\nReviewed-by: Matthieu Baerts (NGI0) \u003cmatttbe@kernel.org\u003e\nSigned-off-by: Matthieu Baerts (NGI0) \u003cmatttbe@kernel.org\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "9415d375d8520e0ed55f0c0b058928da9a5b5b3d", "tree": "848535284c0084d4c6d8d5d4a9cef9bb28eee567", "parents": [ "799a829507506924add8a7620493adc1c3cfda30" ], "author": { "name": "Kuniyuki Iwashima", "email": "kuniyu@amazon.com", "time": "Fri Jul 26 17:19:53 2024 -0700" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Mon Jul 29 11:36:48 2024 +0100" }, "message": "rtnetlink: Don\u0027t ignore IFLA_TARGET_NETNSID when ifname is specified in rtnl_dellink().\n\nThe cited commit accidentally replaced tgt_net with net in rtnl_dellink().\n\nAs a result, IFLA_TARGET_NETNSID is ignored if the interface is specified\nwith IFLA_IFNAME or IFLA_ALT_IFNAME.\n\nLet\u0027s pass tgt_net to rtnl_dev_get().\n\nFixes: cc6090e985d7 (\"net: rtnetlink: introduce helper to get net_device instance by ifname\")\nSigned-off-by: Kuniyuki Iwashima \u003ckuniyu@amazon.com\u003e\nReviewed-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "799a829507506924add8a7620493adc1c3cfda30", "tree": "ed38d7c826b8dc63bd997cc5da06293482d167cb", "parents": [ "05f76b2d634e65ab34472802d9b142ea9e03f74e" ], "author": { "name": "Andy Chiu", "email": "andy.chiu@sifive.com", "time": "Fri Jul 26 15:06:50 2024 +0800" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Mon Jul 29 11:35:37 2024 +0100" }, "message": "net: axienet: start napi before enabling Rx/Tx\n\nsoftirq may get lost if an Rx interrupt comes before we call\nnapi_enable. Move napi_enable in front of axienet_setoptions(), which\nturns on the device, to address the issue.\n\nLink: https://lists.gnu.org/archive/html/qemu-devel/2024-07/msg06160.html\nFixes: cc37610caaf8 (\"net: axienet: implement NAPI and GRO receive\")\nSigned-off-by: Andy Chiu \u003candy.chiu@sifive.com\u003e\nReviewed-by: Eric Dumazet \u003cedumazet@google.com\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "05f76b2d634e65ab34472802d9b142ea9e03f74e", "tree": "dff5d5403fa979a2049f3cb7cf5bb67190d47eb5", "parents": [ "e96a79b19a8499ce2f22ccf0e6b0192e9dcff001" ], "author": { "name": "Subash Abhinov Kasiviswanathan", "email": "quic_subashab@quicinc.com", "time": "Fri Jul 26 13:41:05 2024 -0700" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Mon Jul 29 11:31:50 2024 +0100" }, "message": "tcp: Adjust clamping window for applications specifying SO_RCVBUF\n\ntp-\u003escaling_ratio is not updated based on skb-\u003elen/skb-\u003etruesize once\nSO_RCVBUF is set leading to the maximum window scaling to be 25% of\nrcvbuf after\ncommit dfa2f0483360 (\"tcp: get rid of sysctl_tcp_adv_win_scale\")\nand 50% of rcvbuf after\ncommit 697a6c8cec03 (\"tcp: increase the default TCP scaling ratio\").\n50% tries to emulate the behavior of older kernels using\nsysctl_tcp_adv_win_scale with default value.\n\nSystems which were using a different values of sysctl_tcp_adv_win_scale\nin older kernels ended up seeing reduced download speeds in certain\ncases as covered in https://lists.openwall.net/netdev/2024/05/15/13\nWhile the sysctl scheme is no longer acceptable, the value of 50% is\na bit conservative when the skb-\u003elen/skb-\u003etruesize ratio is later\ndetermined to be ~0.66.\n\nApplications not specifying SO_RCVBUF update the window scaling and\nthe receiver buffer every time data is copied to userspace. This\ncomputation is now used for applications setting SO_RCVBUF to update\nthe maximum window scaling while ensuring that the receive buffer\nis within the application specified limit.\n\nFixes: dfa2f0483360 (\"tcp: get rid of sysctl_tcp_adv_win_scale\")\nSigned-off-by: Sean Tranchetti \u003cquic_stranche@quicinc.com\u003e\nSigned-off-by: Subash Abhinov Kasiviswanathan \u003cquic_subashab@quicinc.com\u003e\nReviewed-by: Eric Dumazet \u003cedumazet@google.com\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "e96a79b19a8499ce2f22ccf0e6b0192e9dcff001", "tree": "308588eaef95369854dbcd29bfc8f37d2675afd3", "parents": [ "9da49aa80d686582bc3a027112a30484c9be6b6e", "0d6ccfe6b319d56da63b7d7cfbcecd92780a680d" ], "author": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Mon Jul 29 10:59:08 2024 +0100" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Mon Jul 29 10:59:08 2024 +0100" }, "message": "Merge branch \u0027ethtool-rss-fixes\u0027 into main\n\nJakub Kicinski says;\n\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\nethtool: more RSS fixes\n\nMore fixes for RSS setting. First two patches fix my own bugs\nin bnxt conversion to the new API. The third patch fixes\nwhat seems to be a 10 year old issue (present since the Linux\nRSS API was created). Fourth patch fixes an issue with\nthe XArray state being out of sync. And then a small test.\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "0d6ccfe6b319d56da63b7d7cfbcecd92780a680d", "tree": "308588eaef95369854dbcd29bfc8f37d2675afd3", "parents": [ "dc9755370e1c5965d16dff98c9877f5b1847e367" ], "author": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Thu Jul 25 15:23:53 2024 -0700" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Mon Jul 29 10:59:08 2024 +0100" }, "message": "selftests: drv-net: rss_ctx: check for all-zero keys\n\nWe had a handful of bugs relating to key being either all 0\nor just reported incorrectly as all 0. Check for this in\nthe tests.\n\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\nReviewed-by: Petr Machata \u003cpetrm@nvidia.com\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "dc9755370e1c5965d16dff98c9877f5b1847e367", "tree": "5409f0de7f389232f84ef7ac80bc5a424cea560b", "parents": [ "7195f0ef7f5b8c678cf28de7c9b619cb908b482c" ], "author": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Thu Jul 25 15:23:52 2024 -0700" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Mon Jul 29 10:59:08 2024 +0100" }, "message": "ethtool: fix the state of additional contexts with old API\n\nWe expect drivers implementing the new create/modify/destroy\nAPI to populate the defaults in struct ethtool_rxfh_context.\nIn legacy API ctx isn\u0027t even passed, and rxfh.indir / rxfh.key\nare NULL so drivers can\u0027t give us defaults even if they want to.\nCall get_rxfh() to fetch the values. We can reuse rxfh_dev\nfor the get_rxfh(), rxfh stores the input from the user.\n\nThis fixes IOCTL reporting 0s instead of the default key /\nindir table for drivers using legacy API.\n\nAdd a check to try to catch drivers using the new API\nbut not populating the key.\n\nFixes: 7964e7884643 (\"net: ethtool: use the tracking array for get_rxfh on custom RSS contexts\")\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\nReviewed-by: Edward Cree \u003cecree.xilinx@gmail.com\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "7195f0ef7f5b8c678cf28de7c9b619cb908b482c", "tree": "feb3ab04fc87456e3c61fcff0dd3092c4edde030", "parents": [ "9dbad38336a9c9a6e77df07c6c770ff6cf55c365" ], "author": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Thu Jul 25 15:23:51 2024 -0700" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Mon Jul 29 10:59:07 2024 +0100" }, "message": "ethtool: fix setting key and resetting indir at once\n\nThe indirection table and the key follow struct ethtool_rxfh\nin user memory.\n\nTo reset the indirection table user space calls SET_RXFH with\ntable of size 0 (OTOH to say \"no change\" it should use -1 / ~0).\nThe logic for calculating the offset where they key sits is\nincorrect in this case, as kernel would still offset by the full\ntable length, while for the reset there is no indir table and\nkey is immediately after the struct.\n\n $ ethtool -X eth0 default hkey 01:02:03...\n $ ethtool -x eth0\n [...]\n RSS hash key:\n00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00\n [...]\n\nFixes: 3de0b592394d (\"ethtool: Support for configurable RSS hash key\")\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "9dbad38336a9c9a6e77df07c6c770ff6cf55c365", "tree": "cde409c88de84dec2a815ccda9c65a0cf83babcb", "parents": [ "daefd348a5938d2256d304b57a9e787a83bb58d9" ], "author": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Thu Jul 25 15:23:50 2024 -0700" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Mon Jul 29 10:59:07 2024 +0100" }, "message": "eth: bnxt: populate defaults in the RSS context struct\n\nAs described in the kdoc for .create_rxfh_context we are responsible\nfor populating the defaults. The core will not call .get_rxfh\nfor non-0 context.\n\nThe problem can be easily observed since Netlink doesn\u0027t currently\nuse the cache. Using netlink ethtool:\n\n $ ethtool -x eth0 context 1\n [...]\n RSS hash key:\n 13:60:cd:60:14:d3:55:36:86:df:90:f2:96:14:e2:21:05:57:a8:8f:a5:12:5e:54:62:7f:fd:3c:15:7e:76:05:71:42:a2:9a:73:80:09:9c\n RSS hash function:\n toeplitz: on\n xor: off\n crc32: off\n\nBut using IOCTL ethtool shows:\n\n $ ./ethtool-old -x eth0 context 1\n [...]\n RSS hash key:\n 00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00\n RSS hash function:\n Operation not supported\n\nFixes: 7964e7884643 (\"net: ethtool: use the tracking array for get_rxfh on custom RSS contexts\")\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\nReviewed-by: Pavan Chebbi \u003cpavan.chebbi@broadcom.com\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "daefd348a5938d2256d304b57a9e787a83bb58d9", "tree": "46dc98451fb6d914fd6dc357ae564523bfb51b04", "parents": [ "9da49aa80d686582bc3a027112a30484c9be6b6e" ], "author": { "name": "Jakub Kicinski", "email": "kuba@kernel.org", "time": "Thu Jul 25 15:23:49 2024 -0700" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Mon Jul 29 10:59:07 2024 +0100" }, "message": "eth: bnxt: reject unsupported hash functions\n\nIn commit under Fixes I split the bnxt_set_rxfh_context() function,\nand attached the appropriate chunks to new ops. I missed that\nbnxt_set_rxfh_context() gets called after some initial checks\nin bnxt_set_rxfh(), namely that the hash function is Toeplitz.\n\nFixes: 5c466b4d4e75 (\"eth: bnxt: move from .set_rxfh to .create_rxfh_context and friends\")\nSigned-off-by: Jakub Kicinski \u003ckuba@kernel.org\u003e\nReviewed-by: Pavan Chebbi \u003cpavan.chebbi@broadcom.com\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "9da49aa80d686582bc3a027112a30484c9be6b6e", "tree": "6ac72bf26a011a6d42b1fbba48233bb86bba59d3", "parents": [ "301927d2d2eb8e541357ba850bc7a1a74dbbd670" ], "author": { "name": "Jeongjun Park", "email": "aha310510@gmail.com", "time": "Fri Jul 26 06:40:49 2024 +0900" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Mon Jul 29 10:55:15 2024 +0100" }, "message": "tun: Add missing bpf_net_ctx_clear() in do_xdp_generic()\n\nThere are cases where do_xdp_generic returns bpf_net_context without\nclearing it. This causes various memory corruptions, so the missing\nbpf_net_ctx_clear must be added.\n\nReported-by: syzbot+44623300f057a28baf1e@syzkaller.appspotmail.com\nFixes: fecef4cd42c6 (\"tun: Assign missing bpf_net_context.\")\nSigned-off-by: Jeongjun Park \u003caha310510@gmail.com\u003e\nAcked-by: Jason Wang \u003cjasowang@redhat.com\u003e\nReviewed-by: Willem de Bruijn \u003cwillemb@google.com\u003e\nReported-by: syzbot+3c2b6d5d4bec3b904933@syzkaller.appspotmail.com\nReported-by: syzbot+707d98c8649695eaf329@syzkaller.appspotmail.com\nReported-by: syzbot+c226757eb784a9da3e8b@syzkaller.appspotmail.com\nReported-by: syzbot+61a1cfc2b6632363d319@syzkaller.appspotmail.com\nReported-by: syzbot+709e4c85c904bcd62735@syzkaller.appspotmail.com\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "dc1c8034e31b14a2e5e212104ec508aec44ce1b9", "tree": "64dded44727b68ac11ae8951597b3cb7d5a2a736", "parents": [ "cb04e8b1d2f24c4c2c92f7b7529031fc35a16fed" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jul 28 20:24:12 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jul 28 20:24:12 2024 -0700" }, "message": "minmax: simplify min()/max()/clamp() implementation\n\nNow that we no longer have any C constant expression contexts (ie array\nsize declarations or static initializers) that use min() or max(), we\ncan simpify the implementation by not having to worry about the result\nstaying as a C constant expression.\n\nSo now we can unconditionally just use temporary variables of the right\ntype, and get rid of the excessive expansion that used to come from the\nuse of\n\n __builtin_choose_expr(__is_constexpr(...), ..\n\nto pick the specialized code for constant expressions.\n\nAnother expansion simplification is to pass the temporary variables (in\naddition to the original expression) to our __types_ok() macro. That\nmay superficially look like it complicates the macro, but when we only\nwant the type of the expression, expanding the temporary variable names\nis much simpler and smaller than expanding the potentially complicated\noriginal expression.\n\nAs a result, on my machine, doing a\n\n $ time make drivers/staging/media/atomisp/pci/isp/kernels/ynr/ynr_1.0/ia_css_ynr.host.i\n\ngoes from\n\n\treal\t0m16.621s\n\tuser\t0m15.360s\n\tsys\t0m1.221s\n\nto\n\n\treal\t0m2.532s\n\tuser\t0m2.091s\n\tsys\t0m0.452s\n\nbecause the token expansion goes down dramatically.\n\nIn particular, the longest line expansion (which was line 71 of that\n\u0027ia_css_ynr.host.c\u0027 file) shrinks from 23,338kB (yes, 23MB for one\nsingle line) to \"just\" 1,444kB (now \"only\" 1.4MB).\n\nAnd yes, that line is still the line from hell, because it\u0027s doing\nmultiple levels of \"min()/max()\" expansion thanks to some of them being\nhidden inside the uDIGIT_FITTING() macro.\n\nLorenzo has a nice cleanup patch that makes that driver use inline\nfunctions instead of macros for sDIGIT_FITTING() and uDIGIT_FITTING(),\nwhich will fix that line once and for all, but the 16-fold reduction in\nthis case does show why we need to simplify these helpers.\n\nCc: David Laight \u003cDavid.Laight@aculab.com\u003e\nCc: Lorenzo Stoakes \u003clorenzo.stoakes@oracle.com\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "cb04e8b1d2f24c4c2c92f7b7529031fc35a16fed", "tree": "78d27543d36ecb38102070bf5e640551f4ebfdae", "parents": [ "9f499b8c791d2983c0a31a543c51d1b2f15e8755" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jul 28 17:32:05 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jul 28 20:23:27 2024 -0700" }, "message": "minmax: don\u0027t use max() in situations that want a C constant expression\n\nWe only had a couple of array[] declarations, and changing them to just\nuse \u0027MAX()\u0027 instead of \u0027max()\u0027 fixes the issue.\n\nThis will allow us to simplify our min/max macros enormously, since they\ncan now unconditionally use temporary variables to avoid using the\nargument values multiple times.\n\nCc: David Laight \u003cDavid.Laight@aculab.com\u003e\nCc: Lorenzo Stoakes \u003clorenzo.stoakes@oracle.com\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "9f499b8c791d2983c0a31a543c51d1b2f15e8755", "tree": "061fcd36991e93a95a0550763b0a560d9a1a6798", "parents": [ "1a251f52cfdc417c84411a056bc142cbd77baef4" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jul 28 17:06:20 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jul 28 17:06:20 2024 -0700" }, "message": "minmax: scsi: fix mis-use of \u0027clamp()\u0027 in sr.c\n\nWhile working on simplifying the minmax functions, and avoiding\nexcessive macro expansion, it turns out that the sr.c use of the\n\u0027clamp()\u0027 macro has the arguments the wrong way around.\n\nThe clamp logic is\n\n\tval \u003d clamp(in, low, high);\n\nand it returns the input clamped to the low/high limits. But sr.c ddid\n\n\tspeed \u003d clamp(0, speed, 0xffff / 177);\n\nwhich clamps the value \u00270\u0027 to the range \u0027[speed, 0xffff / 177]\u0027 and ends\nup being nonsensical.\n\nHappily, I don\u0027t think anybody ever cared.\n\nFixes: 9fad9d560af5 (\"scsi: sr: Fix unintentional arithmetic wraparound\")\nCc: Justin Stitt \u003cjustinstitt@google.com\u003e\nCc: Kees Cook \u003ckeescook@chromium.org\u003e\nCc: Martin K. Petersen \u003cmartin.petersen@oracle.com\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "1a251f52cfdc417c84411a056bc142cbd77baef4", "tree": "2ea60e35ecbb9891dd8746180bf48154095f1272", "parents": [ "8400291e289ee6b2bf9779ff1c83a291501f017b" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jul 28 15:49:18 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jul 28 15:49:18 2024 -0700" }, "message": "minmax: make generic MIN() and MAX() macros available everywhere\n\nThis just standardizes the use of MIN() and MAX() macros, with the very\ntraditional semantics. The goal is to use these for C constant\nexpressions and for top-level / static initializers, and so be able to\nsimplify the min()/max() macros.\n\nThese macro names were used by various kernel code - they are very\ntraditional, after all - and all such users have been fixed up, with a\nfew different approaches:\n\n - trivial duplicated macro definitions have been removed\n\n Note that \u0027trivial\u0027 here means that it\u0027s obviously kernel code that\n already included all the major kernel headers, and thus gets the new\n generic MIN/MAX macros automatically.\n\n - non-trivial duplicated macro definitions are guarded with #ifndef\n\n This is the \"yes, they define their own versions, but no, the include\n situation is not entirely obvious, and maybe they don\u0027t get the\n generic version automatically\" case.\n\n - strange use case #1\n\n A couple of drivers decided that the way they want to describe their\n versioning is with\n\n\t#define MAJ 1\n\t#define MIN 2\n\t#define DRV_VERSION __stringify(MAJ) \".\" __stringify(MIN)\n\n which adds zero value and I just did my Alexander the Great\n impersonation, and rewrote that pointless Gordian knot as\n\n\t#define DRV_VERSION \"1.2\"\n\n instead.\n\n - strange use case #2\n\n A couple of drivers thought that it\u0027s a good idea to have a random\n \u0027MIN\u0027 or \u0027MAX\u0027 define for a value or index into a table, rather than\n the traditional macro that takes arguments.\n\n These values were re-written as C enum\u0027s instead. The new\n function-line macros only expand when followed by an open\n parenthesis, and thus don\u0027t clash with enum use.\n\nHappily, there weren\u0027t really all that many of these cases, and a lot of\nusers already had the pattern of using \u0027#ifndef\u0027 guarding (or in one\ncase just using \u0027#undef MIN\u0027) before defining their own private version\nthat does the same thing. I left such cases alone.\n\nCc: David Laight \u003cDavid.Laight@aculab.com\u003e\nCc: Lorenzo Stoakes \u003clorenzo.stoakes@oracle.com\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "8400291e289ee6b2bf9779ff1c83a291501f017b", "tree": "8f6df59c137acefbebc68a85211e23548791f8eb", "parents": [ "a0c04bd55a467aee3eb647555343ad6971106e86" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jul 28 14:19:55 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jul 28 14:19:55 2024 -0700" }, "message": "Linux 6.11-rc1\n" }, { "commit": "a0c04bd55a467aee3eb647555343ad6971106e86", "tree": "a66abd2bfd726674c5022fdaf3e2c7ec6616a564", "parents": [ "017fa3e89187848fd056af757769c9e66ac3e93d", "3415b10a03945b0da4a635e146750dfe5ce0f448" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jul 28 14:02:48 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jul 28 14:02:48 2024 -0700" }, "message": "Merge tag \u0027kbuild-fixes-v6.11\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild\n\nPull Kbuild fixes from Masahiro Yamada:\n\n - Fix RPM package build error caused by an incorrect locale setup\n\n - Mark modules.weakdep as ghost in RPM package\n\n - Fix the odd combination of -S and -c in stack protector scripts,\n which is an error with the latest Clang\n\n* tag \u0027kbuild-fixes-v6.11\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild:\n kbuild: Fix \u0027-S -c\u0027 in x86 stack protector scripts\n kbuild: rpm-pkg: ghost modules.weakdep file\n kbuild: rpm-pkg: Fix C locale setup\n" }, { "commit": "017fa3e89187848fd056af757769c9e66ac3e93d", "tree": "ce3cd30c5a05f86b0573d19e612d47efa2b27dd9", "parents": [ "4477b39c32fdc03363affef4b11d48391e6dc9ff" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jul 28 13:50:01 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jul 28 13:50:01 2024 -0700" }, "message": "minmax: simplify and clarify min_t()/max_t() implementation\n\nThis simplifies the min_t() and max_t() macros by no longer making them\nwork in the context of a C constant expression.\n\nThat means that you can no longer use them for static initializers or\nfor array sizes in type definitions, but there were only a couple of\nsuch uses, and all of them were converted (famous last words) to use\nMIN_T/MAX_T instead.\n\nCc: David Laight \u003cDavid.Laight@aculab.com\u003e\nCc: Lorenzo Stoakes \u003clorenzo.stoakes@oracle.com\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "4477b39c32fdc03363affef4b11d48391e6dc9ff", "tree": "29fdc4e2d7b47dd23deb73f0660f358669d5c037", "parents": [ "7e2d0ba732ad649c557b49b0a3ef69d54df256d0" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jul 28 13:03:48 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jul 28 13:41:14 2024 -0700" }, "message": "minmax: add a few more MIN_T/MAX_T users\n\nCommit 3a7e02c040b1 (\"minmax: avoid overly complicated constant\nexpressions in VM code\") added the simpler MIN_T/MAX_T macros in order\nto avoid some excessive expansion from the rather complicated regular\nmin/max macros.\n\nThe complexity of those macros stems from two issues:\n\n (a) trying to use them in situations that require a C constant\n expression (in static initializers and for array sizes)\n\n (b) the type sanity checking\n\nand MIN_T/MAX_T avoids both of these issues.\n\nNow, in the whole (long) discussion about all this, it was pointed out\nthat the whole type sanity checking is entirely unnecessary for\nmin_t/max_t which get a fixed type that the comparison is done in.\n\nBut that still leaves min_t/max_t unnecessarily complicated due to\nworries about the C constant expression case.\n\nHowever, it turns out that there really aren\u0027t very many cases that use\nmin_t/max_t for this, and we can just force-convert those.\n\nThis does exactly that.\n\nWhich in turn will then allow for much simpler implementations of\nmin_t()/max_t(). All the usual \"macros in all upper case will evaluate\nthe arguments multiple times\" rules apply.\n\nWe should do all the same things for the regular min/max() vs MIN/MAX()\ncases, but that has the added complexity of various drivers defining\ntheir own local versions of MIN/MAX, so that needs another level of\nfixes first.\n\nLink: https://lore.kernel.org/all/b47fad1d0cf8449886ad148f8c013dae@AcuMS.aculab.com/\nCc: David Laight \u003cDavid.Laight@aculab.com\u003e\nCc: Lorenzo Stoakes \u003clorenzo.stoakes@oracle.com\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "7e2d0ba732ad649c557b49b0a3ef69d54df256d0", "tree": "f6ca333fd542d78c0ebe272ab46a99199766e680", "parents": [ "e172f1e9068807a336c0429b6c57d29bded8d891", "92a286e90203ce3e6c3a6d945fa36da419c3671f" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jul 28 11:51:51 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jul 28 11:51:51 2024 -0700" }, "message": "Merge tag \u0027ubifs-for-linus-6.11-rc1-take2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/rw/ubifs\n\nPull UBI and UBIFS updates from Richard Weinberger:\n\n - Many fixes for power-cut issues by Zhihao Cheng\n\n - Another ubiblock error path fix\n\n - ubiblock section mismatch fix\n\n - Misc fixes all over the place\n\n* tag \u0027ubifs-for-linus-6.11-rc1-take2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/rw/ubifs:\n ubi: Fix ubi_init() ubiblock_exit() section mismatch\n ubifs: add check for crypto_shash_tfm_digest\n ubifs: Fix inconsistent inode size when powercut happens during appendant writing\n ubi: block: fix null-pointer-dereference in ubiblock_create()\n ubifs: fix kernel-doc warnings\n ubifs: correct UBIFS_DFS_DIR_LEN macro definition and improve code clarity\n mtd: ubi: Restore missing cleanup on ubi_init() failure path\n ubifs: dbg_orphan_check: Fix missed key type checking\n ubifs: Fix unattached inode when powercut happens in creating\n ubifs: Fix space leak when powercut happens in linking tmpfile\n ubifs: Move ui-\u003edata initialization after initializing security\n ubifs: Fix adding orphan entry twice for the same inode\n ubifs: Remove insert_dead_orphan from replaying orphan process\n Revert \"ubifs: ubifs_symlink: Fix memleak of inode-\u003ei_link in error path\"\n ubifs: Don\u0027t add xattr inode into orphan area\n ubifs: Fix unattached xattr inode if powercut happens after deleting\n mtd: ubi: avoid expensive do_div() on 32-bit machines\n mtd: ubi: make ubi_class constant\n ubi: eba: properly rollback inside self_check_eba\n" }, { "commit": "3415b10a03945b0da4a635e146750dfe5ce0f448", "tree": "74bb26d62aeafb0865432c7745b742e7e6116e8f", "parents": [ "d01c14074be79e5f5270498f90530a12583fbf7a" ], "author": { "name": "Nathan Chancellor", "email": "nathan@kernel.org", "time": "Fri Jul 26 11:05:00 2024 -0700" }, "committer": { "name": "Masahiro Yamada", "email": "masahiroy@kernel.org", "time": "Mon Jul 29 03:47:00 2024 +0900" }, "message": "kbuild: Fix \u0027-S -c\u0027 in x86 stack protector scripts\n\nAfter a recent change in clang to stop consuming all instances of \u0027-S\u0027\nand \u0027-c\u0027 [1], the stack protector scripts break due to the kernel\u0027s use\nof -Werror\u003dunused-command-line-argument to catch cases where flags are\nnot being properly consumed by the compiler driver:\n\n $ echo | clang -o - -x c - -S -c -Werror\u003dunused-command-line-argument\n clang: error: argument unused during compilation: \u0027-c\u0027 [-Werror,-Wunused-command-line-argument]\n\nThis results in CONFIG_STACKPROTECTOR getting disabled because\nCONFIG_CC_HAS_SANE_STACKPROTECTOR is no longer set.\n\n\u0027-c\u0027 and \u0027-S\u0027 both instruct the compiler to stop at different stages of\nthe pipeline (\u0027-S\u0027 after compiling, \u0027-c\u0027 after assembling), so having\nthem present together in the same command makes little sense. In this\ncase, the test wants to stop before assembling because it is looking at\nthe textual assembly output of the compiler for either \u0027%fs\u0027 or \u0027%gs\u0027,\nso remove \u0027-c\u0027 from the list of arguments to resolve the error.\n\nAll versions of GCC continue to work after this change, along with\nversions of clang that do or do not contain the change mentioned above.\n\nCc: stable@vger.kernel.org\nFixes: 4f7fd4d7a791 (\"[PATCH] Add the -fstack-protector option to the CFLAGS\")\nFixes: 60a5317ff0f4 (\"x86: implement x86_32 stack protector\")\nLink: https://github.com/llvm/llvm-project/commit/6461e537815f7fa68cef06842505353cf5600e9c [1]\nSigned-off-by: Nathan Chancellor \u003cnathan@kernel.org\u003e\nSigned-off-by: Masahiro Yamada \u003cmasahiroy@kernel.org\u003e\n" }, { "commit": "92a286e90203ce3e6c3a6d945fa36da419c3671f", "tree": "7f9632dc40ba72fa5864b2703c5ad8e396cd0b8e", "parents": [ "054fd15984454f031611d6c63675fc578aad0cb1" ], "author": { "name": "Richard Weinberger", "email": "richard@nod.at", "time": "Sat Jul 13 09:35:19 2024 +0200" }, "committer": { "name": "Richard Weinberger", "email": "richard@nod.at", "time": "Sun Jul 28 20:08:25 2024 +0200" }, "message": "ubi: Fix ubi_init() ubiblock_exit() section mismatch\n\nSince ubiblock_exit() is now called from an init function,\nthe __exit section no longer makes sense.\n\nCc: Ben Hutchings \u003cbwh@kernel.org\u003e\nReported-by: kernel test robot \u003clkp@intel.com\u003e\nCloses: https://lore.kernel.org/oe-kbuild-all/202407131403.wZJpd8n2-lkp@intel.com/\nSigned-off-by: Richard Weinberger \u003crichard@nod.at\u003e\nReviewed-by: Zhihao Cheng \u003cchengzhihao1@huawei.com\u003e\n" }, { "commit": "e172f1e9068807a336c0429b6c57d29bded8d891", "tree": "35db59c0a10f197225f6203270a9a1cffceac750", "parents": [ "e62f81bbd24db746c9b1aa29e7b6423211262ac4", "866d2d36b81d7d0e6d91423b6dd9b1bcfd0510dd" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jul 28 10:52:15 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jul 28 10:52:15 2024 -0700" }, "message": "Merge tag \u0027v6.11-merge\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux\n\nPull turbostat updates from Len Brown:\n\n - Enable turbostat extensions to add both perf and PMT (Intel\n Platform Monitoring Technology) counters via the cmdline\n\n - Demonstrate PMT access with built-in support for Meteor Lake\u0027s\n Die C6 counter\n\n* tag \u0027v6.11-merge\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux:\n tools/power turbostat: version 2024.07.26\n tools/power turbostat: Include umask\u003d%x in perf counter\u0027s config\n tools/power turbostat: Document PMT in turbostat.8\n tools/power turbostat: Add MTL\u0027s PMT DC6 builtin counter\n tools/power turbostat: Add early support for PMT counters\n tools/power turbostat: Add selftests for added perf counters\n tools/power turbostat: Add selftests for SMI, APERF and MPERF counters\n tools/power turbostat: Move verbose counter messages to level 2\n tools/power turbostat: Move debug prints from stdout to stderr\n tools/power turbostat: Fix typo in turbostat.8\n tools/power turbostat: Add perf added counter example to turbostat.8\n tools/power turbostat: Fix formatting in turbostat.8\n tools/power turbostat: Extend --add option with perf counters\n tools/power turbostat: Group SMI counter with APERF and MPERF\n tools/power turbostat: Add ZERO_ARRAY for zero initializing builtin array\n tools/power turbostat: Replace enum rapl_source and cstate_source with counter_source\n tools/power turbostat: Remove anonymous union from rapl_counter_info_t\n tools/power/turbostat: Switch to new Intel CPU model defines\n" }, { "commit": "e62f81bbd24db746c9b1aa29e7b6423211262ac4", "tree": "f735031a802722f4ba7b35a2e8ac65379a44c3b6", "parents": [ "7b5d48188942178985cc185a17c754539cb9a4d3", "a0328b397f3339d8d17a6ec356e94b3c110b010c" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jul 28 09:33:28 2024 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jul 28 09:33:28 2024 -0700" }, "message": "Merge tag \u0027cxl-for-6.11\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/cxl/cxl\n\nPull CXL updates from Dave Jiang:\n \"Core:\n\n - A CXL maturity map has been added to the documentation to detail\n the current state of CXL enabling.\n\n It provides the status of the current state of various CXL features\n to inform current and future contributors of where things are and\n which areas need contribution.\n\n - A notifier handler has been added in order for a newly created CXL\n memory region to trigger the abstract distance metrics calculation.\n\n This should bring parity for CXL memory to the same level vs\n hotplugged DRAM for NUMA abstract distance calculation. The\n abstract distance reflects relative performance used for memory\n tiering handling.\n\n - An addition for XOR math has been added to address the CXL DPA to\n SPA translation.\n\n CXL address translation did not support address interleave math\n with XOR prior to this change.\n\n Fixes:\n\n - Fix to address race condition in the CXL memory hotplug notifier\n\n - Add missing MODULE_DESCRIPTION() for CXL modules\n\n - Fix incorrect vendor debug UUID define\n\n Misc:\n\n - A warning has been added to inform users of an unsupported\n configuration when mixing CXL VH and RCH/RCD hierarchies\n\n - The ENXIO error code has been replaced with EBUSY for inject poison\n limit reached via debugfs and cxl-test support\n\n - Moving the PCI config read in cxl_dvsec_rr_decode() to avoid\n unnecessary PCI config reads\n\n - A refactor to a common struct for DRAM and general media CXL\n events\"\n\n* tag \u0027cxl-for-6.11\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/cxl/cxl:\n cxl/core/pci: Move reading of control register to immediately before usage\n cxl: Remove defunct code calculating host bridge target positions\n cxl/region: Verify target positions using the ordered target list\n cxl: Restore XOR\u0027d position bits during address translation\n cxl/core: Fold cxl_trace_hpa() into cxl_dpa_to_hpa()\n cxl/test: Replace ENXIO with EBUSY for inject poison limit reached\n cxl/memdev: Replace ENXIO with EBUSY for inject poison limit reached\n cxl/acpi: Warn on mixed CXL VH and RCH/RCD Hierarchy\n cxl/core: Fix incorrect vendor debug UUID define\n Documentation: CXL Maturity Map\n cxl/region: Simplify cxl_region_nid()\n cxl/region: Support to calculate memory tier abstract distance\n cxl/region: Fix a race condition in memory hotplug notifier\n cxl: add missing MODULE_DESCRIPTION() macros\n cxl/events: Use a common struct for DRAM and General Media events\n" } ], "next": "7b5d48188942178985cc185a17c754539cb9a4d3" }